CN114077712B - Search result ordering method and device - Google Patents

Abstract

The embodiment of the present application discloses a search result sorting method and device, the method comprising: receiving an image search request sent by a client, obtaining search results corresponding to the image search request, and obtaining information clustering results of time-sensitive events; judging whether the image search request is a time-sensitive event search request based on the degree of matching between the search results corresponding to the image search request and the image cluster; when the image search request is a time-sensitive event search request, sorting the search results using the information clustering results to generate a search result queue. By obtaining the real-time information clustering results of time-sensitive events, it is possible to timely judge whether the user's image search request is a time-sensitive event search request. When the image search request is a time-sensitive event search request, the information clustering results are used to sort the search results, and a more accurate search result queue can be obtained.

Description

Search result sorting method and device

Technical Field

The present application relates to the field of Internet technology, and in particular to a search result sorting method and device.

Background Art

When a user performs a search query, corresponding search results will be generated according to the user's search request. When it is determined that the user's search request is a search request for a time-sensitive event, the search results can be adjusted to rank search results with stronger timeliness at a higher position for the user to browse and select.

Currently, there is a certain lag in determining whether a search request is a time-sensitive event, which makes it impossible to promptly feed back search results with strong timeliness to users, resulting in inaccurate sorting of the search result queue corresponding to the user's search request.

Summary of the invention

In view of this, an embodiment of the present application provides a search result sorting method and device, which can judge the user's search request in real time and sort the search results accordingly.

To solve the above problems, the technical solutions provided in the embodiments of the present application are as follows:

A search result sorting method, the method comprising:

Receive a picture search request sent by a client, and obtain search results corresponding to the picture search request;

Acquire information clustering results of time-sensitive events, where the information clustering results of time-sensitive events are obtained by clustering information generated in a preset time period in the network, and the information clustering results include at least one picture cluster, and each of the picture clusters includes at least one picture;

Determining whether the image search request is a timeliness event search request according to a matching degree between the search result corresponding to the image search request and the image cluster;

When the image search request is a time-sensitive event search request, the search results are sorted using the information clustering results to generate a search result queue.

In a possible implementation, judging whether the image search request is a timeliness event search request according to a matching degree between a search result corresponding to the image search request and the image cluster includes:

Determine a first number of images in each of the image clusters that appear in the search results corresponding to the image search request;

Determine the picture clusters whose first number is greater than a first threshold as first target picture clusters;

If the number of first target image clusters to which the search results corresponding to the image search request belong is greater than a second threshold, it is determined that the image search request is a timeliness event search request.

In a possible implementation, when the image search request is a time-sensitive event search request, sorting the search results using the information clustering results to generate a search result queue includes:

When the image search request is a time-sensitive event search request, obtaining cluster features of a second target image cluster to which the search results corresponding to the image search request belong; the second target image cluster is an image cluster corresponding to each image in the search results corresponding to the image search request;

sorting the second target image clusters according to the cluster features of the second target image clusters to generate a sorting result;

Selecting a picture in each of the second target picture clusters as the first target search result;

According to the sorting result, the first target search result and other search results in the search results that do not belong to the second target image cluster are sorted to generate a search result queue, in which the sorting position of the first target search result is higher than the sorting positions of the other search results.

In a possible implementation, obtaining the cluster feature of the second target image cluster to which the search result corresponding to the image search request belongs includes:

Obtain a second number of images in a second target image cluster to which the search results corresponding to the image search request belong, and a number of source websites corresponding to the images in the second target image cluster;

Obtaining text features corresponding to the pictures in the second target picture cluster;

The step of sorting the target image clusters according to the cluster features of the second target image clusters to generate a sorting result includes:

The second target image cluster is sorted according to the matching degree between the query term included in the search request and the text features corresponding to the images in the second target image cluster, the second number of images in the second target image cluster, and the number of source websites corresponding to the images in the second target image cluster to generate a sorting result.

In a possible implementation, the acquiring text features corresponding to the pictures in the second target picture cluster includes:

Obtaining the word frequency of each word in the description text corresponding to the picture in the second target picture cluster, and taking at least one word with the highest word frequency as the text feature corresponding to the picture in the second target picture cluster;

Alternatively, the feature vectors of each word in the description text corresponding to the picture in the second target picture cluster are obtained, and at least one feature vector with the largest number of occurrences in the feature vectors is used as the text feature corresponding to the picture in the second target picture cluster.

In a possible implementation, the method further includes:

Determining a junk information cluster in the information clustering result of the timeliness event, and removing the junk information cluster in the information clustering result of the timeliness event;

The step of determining a spam information cluster from the information clustering result of the timeliness event includes:

If a third number of pictures in the third target picture cluster is greater than a third threshold, or the number of source websites corresponding to the third target picture cluster is less than a fourth threshold, the third target picture cluster is determined as a spam information cluster, and the third target picture cluster is a picture cluster included in the information clustering result of the timeliness event;

Alternatively, if the similarities between the description texts corresponding to the pictures in the third target picture cluster are all less than the fifth threshold, the third target picture cluster is determined to be a spam cluster, and the third target picture cluster is the picture cluster included in the information clustering result of the timeliness event.

A search result sorting device, the device comprising:

A search result acquisition unit, configured to receive an image search request sent by a client and acquire search results corresponding to the image search request;

An information clustering result acquisition unit, configured to acquire an information clustering result of a time-sensitive event, wherein the information clustering result of the time-sensitive event is obtained by clustering information generated in a preset time period in the network, and the information clustering result includes at least one picture cluster, and each of the picture clusters includes at least one picture;

A judging unit, configured to judge whether the image search request is a timeliness event search request according to a matching degree between the search result corresponding to the image search request and the image cluster;

A search result queue generating unit is used to sort the search results using the information clustering results to generate a search result queue when the image search request is a time-sensitive event search request.

In a possible implementation, the determining unit includes:

A first quantity determination module, configured to determine a first quantity of images in each of the image clusters that appear in the search results corresponding to the image search request;

A picture cluster determination module, configured to determine picture clusters whose first number is greater than a first threshold as first target picture clusters;

The first judgment module is configured to judge that the image search request is a timeliness event search request if the number of first target image clusters to which the search results corresponding to the image search request belong is greater than a second threshold.

In a possible implementation, the search result queue generating unit includes:

A cluster feature acquisition subunit is used to acquire, when the image search request is a time-sensitive event search request, a cluster feature of a second target image cluster to which the search result corresponding to the image search request belongs; the second target image cluster is an image cluster corresponding to each image in the search result corresponding to the image search request;

a sorting result generating subunit, configured to sort the second target image clusters according to the cluster features of the second target image clusters to generate a sorting result;

A first selection subunit, configured to select a picture in each of the second target picture clusters as a first target search result;

The first search result queue generating subunit is used to sort the first target search result and other search results in the search results that do not belong to the second target image cluster according to the sorting result, and generate a search result queue, in which the sorting position of the first target search result is higher than the sorting position of the other search results.

In a possible implementation, the cluster feature acquisition subunit includes:

A source website quantity acquisition module, used to acquire a second quantity of images in a second target image cluster to which the search result corresponding to the image search request belongs, and a quantity of source websites corresponding to the images in the second target image cluster;

A text feature acquisition module, used to acquire text features corresponding to the pictures in the second target picture cluster;

The sorting result generating subunit is specifically used for:

The second target image cluster is sorted according to the matching degree between the query term included in the search request and the text features corresponding to the images in the second target image cluster, the second number of images in the second target image cluster, and the number of source websites corresponding to the images in the second target image cluster to generate a sorting result.

In a possible implementation, the text feature acquisition module is specifically used to:

Obtaining the word frequency of each word in the description text corresponding to the picture in the second target picture cluster, and taking at least one word with the highest word frequency as the text feature corresponding to the picture in the second target picture cluster;

Alternatively, the feature vectors of each word in the description text corresponding to the picture in the second target picture cluster are obtained, and at least one feature vector with the largest number of occurrences in the feature vectors is used as the text feature corresponding to the picture in the second target picture cluster.

In a possible implementation manner, the device further includes:

A spam information cluster determination unit, configured to determine a spam information cluster from the information clustering result of the timeliness event;

A spam information cluster removal unit, used to remove the spam information cluster from the information clustering result of the timeliness event;

The junk information cluster determination unit is specifically used to:

If a third number of pictures in the third target picture cluster is greater than a third threshold, or the number of source websites corresponding to the third target picture cluster is less than a fourth threshold, the third target picture cluster is determined as a spam information cluster, and the third target picture cluster is a picture cluster included in the information clustering result of the timeliness event;

Alternatively, if the similarities between the description texts corresponding to the pictures in the third target picture cluster are all less than the fifth threshold, the third target picture cluster is determined to be a spam cluster, and the third target picture cluster is the picture cluster included in the information clustering result of the timeliness event.

A search result ranking device includes a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors. The one or more programs include instructions for performing the following operations:

Receive a picture search request sent by a client, and obtain search results corresponding to the picture search request;

Acquire information clustering results of time-sensitive events, where the information clustering results of time-sensitive events are obtained by clustering information generated in a preset time period in the network, and the information clustering results include at least one picture cluster, and each of the picture clusters includes at least one picture;

Determining whether the image search request is a timeliness event search request according to a matching degree between the search result corresponding to the image search request and the image cluster;

When the image search request is a time-sensitive event search request, the search results are sorted using the information clustering results to generate a search result queue.

A computer-readable medium stores instructions which, when executed by one or more processors, enable a device to execute the search result ranking method.

It can be seen that the embodiments of the present application have the following beneficial effects:

In the search result sorting method provided in the embodiment of the present application, the server receives the image search request sent by the client, obtains the search result corresponding to the image search request, and obtains the information clustering result of the timeliness event, wherein the information clustering result of the timeliness event is obtained by clustering the information generated within a preset time period in the network, and the information clustering result includes at least one image cluster, and each image cluster includes at least one image; the server determines whether the image search request is a timeliness event search request according to the matching degree between the search result corresponding to the image search request and the image cluster; when the image search request is a timeliness event search request, the search results are sorted using the information clustering result to generate a search result queue. Among them, according to the matching degree between the search result corresponding to the image search request and the image cluster, the image search request belonging to the timeliness event search request is determined to be real-time. Through the matching degree between the search result corresponding to the image search request and the image cluster, it is possible to timely find out whether the user's image search request is a timeliness event search request. When the image search request is a timeliness event search request, the search results are sorted using the information clustering result, and a more accurate search result queue can be obtained.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of a framework of an exemplary application scenario provided in an embodiment of the present application;

FIG2 is a flow chart of a search result sorting method provided in an embodiment of the present application;

FIG3 is a flow chart of a method for determining whether a search request is a time-sensitive event search request provided by an embodiment of the present application;

FIG4 is a flowchart of a method for generating a search result queue provided in an embodiment of the present application;

FIG5 is a schematic diagram of the structure of a search result sorting device provided in an embodiment of the present application;

FIG6 is a schematic diagram of the structure of a device for sorting search results provided in an embodiment of the present application;

FIG. 7 is a schematic diagram of the structure of a server device provided in an embodiment of the present application.

DETAILED DESCRIPTION

In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

In order to facilitate the understanding and explanation of the technical solutions provided by the embodiments of the present application, the background technology of the present application will be described below.

After studying the traditional search result sorting, the inventor found that before sorting the search results, it is first determined whether the search request is a time-sensitive event search request, where a time-sensitive event refers to a hot event that occurred recently. In the prior art, time-sensitive events are determined by obtaining the frequency of occurrence of query terms in search requests. When the frequency of occurrence of query terms in search requests on the network is high, the event corresponding to the query term is determined as a time-sensitive event. However, by determining whether the event corresponding to the query term is a time-sensitive event based on the frequency of occurrence of the query term in the search request, a large number of search requests with the query term are required to meet the condition of determining the event as a time-sensitive event, which makes the determination of time-sensitive events have a certain lag, and it is not possible to promptly determine whether the search request is a time-sensitive event search request when the search request is made.

Based on this, an embodiment of the present application provides a search result sorting method, which includes: first, receiving an image search request sent by a client, obtaining a search result corresponding to the image search request, and also obtaining an information clustering result of a timeliness event, wherein the information clustering result of the timeliness event is obtained by clustering information generated within a preset time period in the network, and the information clustering result includes at least one image cluster, and each image cluster includes at least one image; secondly, judging whether the image search request is a timeliness event search request based on the degree of matching between the search result corresponding to the image search request and the image cluster, thereby realizing real-time judgment of the image search request as a timeliness event search request; finally, when the image search request is a timeliness event search request, the information clustering result is used to sort the search results, and a search result queue is generated, so as to obtain more accurate search results.

See FIG1 , which is a schematic diagram of a framework of an exemplary application scenario provided by an embodiment of the present application. The search result sorting method provided by an embodiment of the present application can be applied to a server 20 .

In actual applications, the server 20 obtains the search request sent by the client 10, obtains the corresponding image search results according to the image search request, obtains the information clustering results of the timeliness event, and uses the degree of matching between the search results corresponding to the image search request and the image cluster to determine whether the image search request is a timeliness event search request. When the image search request is a timeliness event search request, the information clustering results are used to sort the search results, generate a search result queue, and send the search result queue to the client 10 so that the client 10 can display the search result queue.

Those skilled in the art will appreciate that the framework diagram shown in FIG1 is only an example in which the embodiments of the present application can be implemented. The scope of application of the embodiments of the present application is not limited in any aspect by the framework.

It should be noted that the client 10 can be any user device that is currently available, under development, or will be developed in the future, and can interact with each other through any form of wired and/or wireless connection (e.g., Wi-Fi, LAN, cellular, coaxial cable, etc.), including but not limited to: smart wearable devices, smart phones, non-smart phones, tablet computers, laptop personal computers, desktop personal computers, minicomputers, mid-range computers, mainframe computers, etc. that are currently available, under development, or will be developed in the future. The implementation of the present application is not limited in this regard. It should also be noted that the server 20 in the embodiment of the present application can be an example of a device that is currently available, under development, or will be developed in the future, and can provide information recommendation application services to users. The implementation of the present application is not limited in this regard.

To facilitate understanding of the technical solution provided in the embodiment of the present application, the search result sorting method provided in the embodiment of the present application will be described below with reference to the accompanying drawings.

See FIG. 2 , which is a flow chart of a search result sorting method provided in an embodiment of the present application. As shown in FIG. 2 , the method may include S201-S204:

S201: Receive an image search request sent by a client, and obtain search results corresponding to the image search request.

The image search request sent by the client may be a request for performing an image search including a query word, and the image search request may be used to indicate an event for performing an image search.

Through the received image search request, the search results corresponding to the image search request can be determined through the information captured in the network. It is understandable that before sending the search results to the client, the order of the search results needs to be adjusted to obtain the search result queue so that the order in which the search results are displayed is more convenient for users to browse and use.

In the embodiment of the present application, by receiving the image search request sent by the client, the search results corresponding to the image search request can be obtained, and the image search request can be promptly judged whether it is a time-sensitive event search request. When the image search request is a time-sensitive event search request, the search results are further sorted to obtain a more accurate sorting result queue.

S202: Obtain information clustering results of time-sensitive events, where the information clustering results of time-sensitive events are obtained by clustering information generated within a preset time period in the network, and the information clustering results include at least one picture cluster, and each picture cluster includes at least one picture.

Time-sensitive events refer to hot events that occurred recently. Time-sensitive events usually first have relevant event information reported on the network, and then attract the attention of users through the dissemination of information. Users further use the client to search for time-sensitive events by sending search requests. Time-sensitive events usually first have relevant information reported, and then attract the attention of users to search for relevant information. It can be seen that by obtaining information generated in a preset time period in the network and clustering the information generated in the preset time period in the network, the information clustering results corresponding to the information generated in the network in the time period can be obtained in a timely manner. The information clustering results are the information clustering results of time-sensitive events.

The specific length of the preset time period is not limited in the embodiment of the present application, and the preset time period can be set accordingly according to the timeliness of the timeliness event to be obtained. For example, when the timeliness of the timeliness event to be obtained is high, the length of the corresponding preset time period is short, such as 3 hours; when the timeliness of the timeliness event to be obtained is slightly lower, the length of the corresponding preset time period can be slightly longer, such as 24 hours.

The information clustering result is obtained by clustering the information generated within a preset time period in the acquired network. The information clustering result includes at least one information cluster, and an information cluster can correspond to the information of a time-sensitive event. An information cluster contains related information of the same type belonging to the same event, for example, pictures of the same type related to the same time-sensitive event or text keywords of the same type related to the same time-sensitive event. Specifically, when the information cluster is a picture cluster, the corresponding information clustering result includes at least one picture cluster. Each picture cluster includes at least one picture, and a picture cluster contains the same pictures. The same picture can be a picture with the same content and the same angle. Pictures obtained by zooming in and out and adding watermarks can be considered to be the same picture.

S203: judging whether the image search request is a timeliness event search request according to the matching degree between the search results corresponding to the image search request and the image cluster.

After obtaining the information clustering results of the timeliness events, the timeliness event search request can be judged according to the information clustering results of the timeliness events.

Specifically, the image search request corresponds to a specific event. By obtaining the image clusters in the information clustering results of the time-sensitive event, it is possible to determine whether the image search request is a time-sensitive event search request based on the matching degree between the search results and the image clusters. When the image search request is a time-sensitive event search request, it is necessary to sort the corresponding search results, adjust the sorting of search results with higher timeliness, and obtain a more accurate search result queue.

According to the matching degree between the search results corresponding to the image search request and the image cluster, the similarity or identity degree between the images in the search results and the images in the image cluster can be obtained. Specifically, when the matching degree between the search results corresponding to the image search request and the image cluster is high, it means that the search results corresponding to the image search request have more images that are similar or identical to the images in the image cluster, and the corresponding search request is a time-sensitive event search request; when the matching degree between the search results corresponding to the image search request and the image cluster is low, it means that the search results corresponding to the image search request have fewer images that are similar or identical to the images in the image cluster, and the corresponding search request is not a time-sensitive event search request.

The present application embodiment provides a specific implementation of S203, please see below.

It should be noted that after obtaining the information clustering results of the timeliness event, the information clustering results need to be loaded in a timely manner so as to judge the image search request in real time through the information clustering results.

In the embodiment of the present application, since the information clustering result is obtained by clustering the information generated in the network in a preset time period, the obtained information clustering result is real-time. By using the image clusters in the information clustering result to judge the image search request, the image search request can be judged more timely so as to accurately sort the search results.

S204: When the image search request is a time-sensitive event search request, the search results are sorted using the information clustering results to generate a search result queue.

The search results obtained from the image search request are not sorted. Before displaying the search results, the search results need to be sorted for display. If the image search request is a time-sensitive event search request, the search results with a high relevance to the time-sensitive event search request and a high timeliness need to be ranked higher so that users can quickly query the search results with a high relevance to the time-sensitive event search request, and avoid the search results with a low relevance to the time-sensitive event being ranked higher in the search result queue, which prevents users from quickly obtaining accurate information.

The obtained search result queue has search results arranged in a certain order. The search result queue can be used for displaying the search results on the client so that the user can browse the search results.

In the embodiment of the present application, when the image search request is a time-sensitive event search request, the search results are sorted using the information clustering results to obtain a search result queue. The search result queue obtained by sorting the information clustering results can more accurately place search results with a high degree of relevance to time-sensitive events and high timeliness at a higher position in the queue, making it easier for users to quickly obtain accurate information.

In addition, in a possible implementation, an embodiment of the present application provides a method for sorting search results using information clustering results to generate a search result queue. For details, please refer to the specific implementation method below.

Based on the relevant contents of S201 to S204 above, it can be known that in an embodiment of the present application, an image search request sent by a client is received, a corresponding search result is obtained, and an information clustering result of a time-sensitive event is obtained. Since the information clustering result of the time-sensitive event obtained is real-time, the search request can be judged in a timely manner using the information clustering result of the time-sensitive event. If the search request is a time-sensitive event search request, the search results are sorted using the information clustering result to generate a search result queue. By clustering the information generated within a preset time period in the network, the information clustering result of the time-sensitive event can be obtained in a timely manner, and the search request can be judged in a timely manner, thereby generating a more accurate search result queue.

It is understandable that not all information generated in the network within a preset time period is related to time-sensitive events, which may result in the existence of spam information clusters in the information clustering results of time-sensitive events that cannot be used for search request judgment and search result sorting. If there are spam information clusters in the information clustering results of time-sensitive events, search requests that are not time-sensitive events may be misjudged as time-sensitive event search requests, affecting the accuracy of the judgment of the search request.

Based on the problem that there are spam information clusters in the obtained information clustering results of timeliness events, which affects the judgment of search requests and the sorting of search results, the embodiment of the present application also provides another implementation method for sorting search results. In this implementation method, in addition to including the above S201 to S204, after executing the above S202, the search result sorting method may also include the following steps:

The junk information clusters are determined in the information clustering results of the time-sensitive events, and the junk information clusters are removed in the information clustering results of the time-sensitive events.

Spam clusters refer to information clusters that affect the judgment of search requests and the ranking of search results. Spam clusters may include information clusters that do not belong to timeliness events, information clusters that are not highly associated with timeliness events, and information clusters that are not highly credible. Among them, information clusters that do not belong to timeliness events refer to information clusters obtained by clustering commonly used information generated within a preset time period on the Internet or publicity information that is not related to specific events. Specifically, for example, it can be information clusters related to common phrases and common expressions, and information clusters related to advertisements. Information clusters that are not highly associated with timeliness events refer to information clusters obtained by clustering information that is partially related to timeliness events. For example, if a timeliness event is mainly composed of people and places, then only the information for the person or the place is partial information that is not highly associated with the validity event, and cannot fully reflect the timeliness event. The information cluster obtained by clustering is an information cluster that is not highly associated with the timeliness event. In addition, information clusters with low credibility may also be spam information clusters. For example, there is a large amount of information in the information cluster, but the number of source websites of the information is small. The information cluster obtained by clustering such information is an information cluster with low credibility.

After determining the spam cluster, the determined spam cluster can be removed from the information clustering result of the timeliness event to obtain an updated information clustering result of the timeliness event. The updated information clustering result can be used to more accurately judge the search request, avoid the influence of the spam cluster on the misjudgment of the search request, and thus improve the accuracy of the search result queue sorting.

In the embodiment of the present application, the spam information cluster may be a spam picture cluster. Then, the spam information cluster is determined in the information clustering result of the timeliness event, and specifically may include A1-A3:

A1: If the third number of pictures in the third target picture cluster is greater than a third threshold, the third target picture cluster is determined as a spam cluster.

The third target picture cluster is a picture cluster included in the information clustering result of the timeliness event. Whether the third target picture cluster is a spam information cluster is determined by judging whether the third number of pictures in the third target picture cluster is greater than the third threshold. Among them, the third threshold can be a larger value, for example, the third threshold is 1000. It can be understood that when the third number of pictures in the third target picture cluster is larger, the corresponding third target picture cluster has more pictures and may not have relevant information for the timeliness event. The corresponding third target picture cluster may be a picture cluster that does not belong to the timeliness event or a picture cluster that is not highly associated with the timeliness event. For example, a picture cluster composed of pictures corresponding to common phrases or a picture cluster composed of propaganda pictures has a large number of pictures, which is irrelevant to the timeliness event and does not include information about the timeliness event. In addition, the third target picture cluster can also be a picture cluster that is not highly associated with the timeliness event. Since the information corresponding to the pictures in this type of picture cluster includes a larger range, the corresponding number of pictures is larger.

A2: If the number of source websites corresponding to the third target image cluster is less than a fourth threshold, the third target image cluster is determined to be a spam cluster.

The third target image cluster is an image cluster included in the information clustering result of the timeliness event. When the number of source websites corresponding to the third target image cluster is less than the fourth threshold, it can be considered that the source range of the images in the third target image cluster is small and the credibility is not high. The fourth threshold can be a small value.

A3: If the similarities between the description texts corresponding to the pictures in the third target picture cluster are all less than the fifth threshold, the third target picture cluster is determined to be a spam cluster.

The third target image cluster is an image cluster included in the information clustering result of the timeliness event. The description text corresponding to the image can characterize the characteristics of the image, and whether the third target image cluster is a spam information cluster can be judged based on the similarity between the description texts corresponding to the images. When the similarity between the description texts corresponding to each image in the third target image cluster is less than the fifth threshold, it means that the similarity between the description texts of each image is low, indicating that the description texts corresponding to each image in the third target image cluster cannot reflect the common characteristics of the images in the cluster, and the image may not have a strong connection with the description text, and may be a more commonly used image, such as a decorative image in a web page or a commonly used emoticon image. The image cluster composed of such images has nothing to do with the specific timeliness event and belongs to a spam information cluster.

Based on the above content, it can be known that in an embodiment of the present application, by removing spam information clusters from the information clustering results of the acquired timeliness events, the influence of spam information clusters on the judgment of image search requests using the information clustering results can be avoided, and it can be more accurately judged whether the image search request is a validity event search request, thereby obtaining a more accurate search result queue.

In a possible implementation, the degree of matching between the search results corresponding to the image search request and the image clusters can be determined based on the number of images in each image cluster in the search results corresponding to the image search request, thereby determining whether the search request is a time-sensitive event search request. Referring to FIG. 3 , which is a flow chart of a method for determining whether a search request is a time-sensitive event search request provided in an embodiment of the present application, as shown in FIG. 3 , the method may include S301-S303:

S301: Determine a first number of images in each image cluster that appear in search results corresponding to an image search request.

The search results corresponding to the image search request are obtained, and the first number of images in each image cluster that appear in the search results is determined. The first number refers to the number of images in the image cluster that appear in the search results. By obtaining the first number of images in each image cluster that appear in the search results, the appearance of images in the image cluster in the search results corresponding to the image search request can be obtained, so as to make a judgment on the search request.

Since the number of images in the search results corresponding to the image search request may be large, the search results corresponding to the image search request may be preliminarily sorted first, and the images with higher sorting are used as the images used to determine the first number of images in each image cluster that appear in the search results. In the embodiments of the present application, the sorting method of the preliminary sorting is not limited, and the preliminary sorting of the images in the search results may be performed based on the relevance of the images in the search results to the search request, the timeliness of the images, the quality of the images, etc.

S302: Determine picture clusters whose first number is greater than a first threshold as first target picture clusters.

After obtaining the first number of images in each image cluster that appear in the search results corresponding to the image search request, the image cluster whose first number is greater than the first threshold is determined as the first target image cluster. The first threshold may be a threshold corresponding to the number of images in the image cluster that appear in the search results, or a threshold corresponding to the ratio of the number of images in the image cluster that appear in the search results to the total number of images in the image cluster. When the first number of image clusters is greater than the first threshold, the number of images or the proportion of images in the image cluster that appear in the search results is high, and the image cluster may be determined as the first target image cluster.

The first target image cluster can be used to further determine whether the search request is a validity event search request. It is understandable that when a picture cluster contains the same pictures, and the number of pictures or the proportion of pictures in the picture cluster that appear in the search results is high, it can be considered that the picture cluster has a high relevance to the search results, and the first target image cluster can be used to further determine whether the search request is a validity event search request. Correspondingly, if a picture cluster does not belong to the first target picture cluster, it means that the number of pictures or the proportion of pictures in the picture cluster that appear in the search results is low, and the relevance to the search results is low. When using the picture cluster to determine whether the search request is a validity event search request, the picture cluster can be ignored.

In an embodiment of the present application, by comparing the first number with the first threshold, the picture cluster whose first number is greater than the first threshold is determined as the first target picture cluster. When judging the image search results, the judgment can be made according to the number of the first target picture clusters. In this way, the picture clusters with a small first number of pictures that appear in the search results corresponding to the picture search request can be excluded, thereby avoiding the influence of such picture clusters on the search request judgment, and improving the accuracy of the search request judgment.

S303: If the number of first target image clusters to which the search results corresponding to the image search request belong is greater than a second threshold, it is determined that the image search request is a timeliness event search request.

When the number of first target image clusters to which the search results corresponding to the image search request belong is large, the search results corresponding to the image search request include images corresponding to multiple time-sensitive events, thereby determining the search request as a time-sensitive event search request. The second threshold may be a threshold corresponding to the number of first target image clusters.

Based on the contents of S301 to S303 above, it can be known that in the embodiment of the present application, based on the images in the search results and the images in the image clusters, the first target image cluster is first determined, and then the number of first target image clusters is used to determine whether the search request is a search request for a timeliness event. The first target image cluster is an image cluster with a high relevance to the search results. When there are a large number of first target image clusters, the search results can be considered to be search results for validity events, and the corresponding search request is a validity event search request. By determining the first target image cluster, the influence of image clusters with a low relevance to the search results can be removed when judging the search request, further improving the accuracy of judging the image search request.

Further, the embodiment of the present application provides a method for sorting search results and generating a search result queue by using information clustering results when the search request is an image search request. Referring to FIG. 4 , this figure is a flow chart of a method for generating a search result queue provided by an embodiment of the present application. As shown in FIG. 4 , the method may include S401-S404:

S401: When the image search request is a time-sensitive event search request, obtain cluster features of a second target image cluster to which the search results corresponding to the image search request belong; the second target image cluster is the image cluster corresponding to each image in the search results corresponding to the image search request.

After determining that the image search request is a validity event search request, the search results need to be sorted. Since the search results may contain search results that are related to the search request for timeliness events but have lower timeliness, the search results with higher timeliness need to be ranked higher to facilitate user browsing and selection.

Before sorting the search results, the cluster features of the second target image cluster of the search results corresponding to the image search request can be obtained. The second target image cluster refers to the image cluster corresponding to each image in the search results corresponding to the image search request, that is, if the images in the image cluster appear in the search results, the image cluster is determined as the second target image cluster.

The cluster feature of the second target image cluster may be a feature used to characterize information related to images in the second target image cluster. The cluster feature of the second target image cluster may be determined based on all or part of the images in the second target image cluster, thereby avoiding the problem of inaccurate cluster features caused by obtaining the cluster feature of the second target image cluster corresponding to the image based on only a single image.

In a possible implementation, the cluster feature of the second target image cluster may include the second number of images in the second target image cluster, the number of corresponding source websites, and the corresponding text features.

Correspondingly, the cluster features of the second target image cluster to which the search results corresponding to the image search request belong are obtained, which may include B1-B2:

B1: Obtain the second number of images in the second target image cluster to which the search results corresponding to the image search request belong, and the number of source websites corresponding to the images in the second target image cluster.

The second number of pictures in the second target picture cluster may be the number of pictures in the second target picture cluster, and the number of source websites corresponding to the pictures in the second target picture cluster refers to the number of source websites corresponding to the pictures in the second target picture cluster.

B2: Obtain text features corresponding to the images in the second target image cluster.

The text features corresponding to the pictures in the second target picture cluster refer to the text features of the description text corresponding to the pictures. It can be understood that the text corresponding to the pictures has features corresponding to the pictures, and the text features corresponding to the pictures in the second target picture cluster can be used as one of the cluster features of the second target picture cluster. In addition, the embodiment of the present application also provides an implementation method for obtaining the text features corresponding to the pictures in the second target picture cluster, please refer to the specific implementation method below.

S402: Sort the second target image clusters according to the cluster features of the second target image clusters to generate a sorting result.

According to the obtained cluster features of the second target image clusters, the second target image clusters are sorted to obtain the sorting results of the second target image clusters. For different second target image clusters, the quality of the image clusters is different. The second target image clusters can be sorted first to obtain the sorting results of the second target image clusters, so as to sort the search results according to the sorting results of the second target image clusters.

In a possible implementation, when the cluster feature of the second target image cluster includes the second number of images in the second target image cluster, the number of corresponding source websites, and the corresponding text feature, sorting the target image clusters according to the cluster feature of the second target image cluster to generate a sorting result may include:

According to the matching degree between the query words included in the search request and the text features corresponding to the pictures in the second target picture cluster, the second number of pictures in the second target picture cluster, and the number of source websites corresponding to the pictures in the second target picture cluster, the second target picture cluster is sorted to generate a sorting result. It should be noted that, corresponding to different query methods, the query words may have different sources. In a possible implementation, if the search request input by the user is a string composed of characters, the query word may be a string or a word segmentation based on the string. If the user inputs a picture, the query word may be a corresponding semantic word obtained after the picture is recognized.

It is understandable that the search request contains query words, the search results are obtained based on the query words, the query words can represent the characteristics of the search results, and the degree of matching between the query words included in the search request and the text features corresponding to the pictures in the second target picture cluster can be used as a reference for sorting. In the embodiment of the present application, the method for determining the degree of matching between the query words included in the search request and the text features corresponding to the pictures in the second target picture cluster is not limited. The feature vector of the query words can be obtained and compared with the text features of the pictures to obtain the degree of matching between the query words and the text features corresponding to the pictures; the query words can also be compared with the text features of the pictures to obtain the degree of matching between the query words and the text features corresponding to the pictures.

When the text features of the pictures in the second target picture cluster have a high degree of matching with the query words, the second number of pictures is large, and the number of source websites corresponding to the pictures is large, the second target picture cluster has a high degree of association with the search request, has richer pictures, and has a high credibility. Therefore, when sorting the second target picture clusters, the second target picture clusters with a high degree of matching, a large second number of pictures, and a large number of source websites corresponding to the pictures can be placed at a higher position in the sorting result, so that the search result queue obtained by sorting according to the sorting result of the second target picture cluster is more accurate.

S403: Selecting a picture in each second target picture cluster as the first target search result.

The second target picture cluster has at least one picture, and one picture can be selected from each second target picture cluster as the first target search result to generate a search result queue. When selecting pictures from the second target picture cluster, the picture with the best picture quality in the second target picture cluster can be selected as the corresponding first target search result. By selecting one picture from each second target picture cluster as the first target search result, one second target picture cluster corresponds to one first target search result, which can avoid duplication of search results, improve the effectiveness of search results in the search result queue, and make the search result queue more accurate.

S404: Sort the first target search result and other search results in the search results that do not belong to the second target image cluster according to the sorting result to generate a search result queue, in which the sorting position of the first target search result is higher than the sorting positions of other search results.

It is understandable that the first target search result is selected from the second target image cluster, has a higher timeliness, and has a higher degree of relevance to the search results of the timeliness event. When sorting the search results, the first target search result can be ranked higher for users to browse and select. The search results corresponding to the search request also have search results that do not correspond to the image cluster. Such search results may be search results with lower timeliness or search results with lower relevance to the search request. When sorting the search results, this type of search results is ranked at a later position.

In some cases, for example, the information clustering results may include image clusters obtained by clustering historical images. Such image clusters are obtained by clustering image information generated within a preset time period, but the images in the image clusters are the same as historical images. Such image clusters can be determined as historical image clusters. When sorting the search results, other search results that do not belong to the second target image cluster can be arranged after the first target search results. For example, the first target search result belonging to the historical image cluster is arranged at a lower position to avoid interference of historical images in the sorting of search results.

According to the obtained sorting result of the second target image cluster, the search results are sorted to generate a search result queue. The search results include the first target search results and other search results, and the first target search results are sorted before the other search results, so that the user can quickly browse the search results with a high degree of relevance to the search request according to the search result queue.

Based on the contents of S401-S404 above, it can be known that in the embodiment of the present application, when the search request is an image search request and a validity event search request, the second target image cluster is sorted using the cluster characteristics of the second target image cluster to which the search results belong, and then an image is selected from the second target image cluster as the first target search result, and finally the search results are sorted according to the sorting result and the first target search result. By sorting the second target image cluster and selecting the image, a first target search result with a high timeliness and a high correlation with the timeliness event can be obtained, and the first target search result is arranged at a higher position in the search result queue, so that a more accurate search result queue can be obtained, so that the user can quickly obtain a search result with a high correlation with the validity event search request.

Based on the above-mentioned content of obtaining text features corresponding to pictures in the second target picture cluster, the embodiment of the present application also provides a method for obtaining text features corresponding to pictures in the second target picture cluster, including two methods C1 and C2 for obtaining text features.

C1: Obtain the word frequency of each word in the description text corresponding to the picture in the second target picture cluster, and use at least one word with the highest word frequency as the text feature corresponding to the picture in the second target picture cluster.

It is understandable that the description text corresponding to the picture in the second target picture cluster can be a sentence composed of multiple words, and the search request usually contains query words, so the description text corresponding to the picture in the second target picture cluster is segmented, and the segmented words are selected as the text features corresponding to the pictures, so that the selected segmented words can be matched with the query words later to determine the matching degree between the text features and the query words.

It should be noted that the description text corresponding to the pictures in the second target picture cluster can be the description text corresponding to all pictures in the second target picture cluster or the description text corresponding to most pictures. The accuracy of the text features can be ensured by obtaining the text features through the description text corresponding to the pictures with a relatively large number. The word frequency of each word segment in the description text is obtained, and the word segment with the highest word frequency can represent the text feature corresponding to the picture. In addition, the text feature corresponding to some images is composed of multiple word segments. Correspondingly, at least one word segment with the highest word frequency can be used as the text feature corresponding to the picture in the second target picture cluster to ensure the integrity of the text feature.

In the above method of selecting word segmentation as text feature based on word frequency, only the frequency of word segmentation is considered, and the influence of semantics on text features is ignored. To solve the above problem, the embodiment of the present application also provides another method for obtaining text features.

C2: Obtain feature vectors of each word in the description text corresponding to the picture in the second target picture cluster, and use at least one feature vector with the largest number of occurrences in the feature vectors as the text feature corresponding to the picture in the second target picture cluster.

On the basis of word segmentation of the description text corresponding to the picture in the second target picture cluster, the feature vectors of each word segmentation are further obtained, and at least one feature vector that appears most times in the feature vector is used as a text feature. By using at least one feature vector that appears most times in the feature vector as a text feature, the feature vector can be determined from two aspects of semantics and frequency, so that the obtained feature vector can better reflect the characteristics of the picture.

Furthermore, the matching degree between the query word included in the search request obtained subsequently and the text features corresponding to the pictures in the second target picture cluster corresponds to the matching degree between the feature vector of the query word and the text features corresponding to the pictures.

Based on the content of obtaining the text features corresponding to the pictures in the second target picture cluster provided in the above-mentioned embodiment of the present application, it can be known that the embodiment of the present application can obtain more accurate text features corresponding to the pictures by taking the most frequent word segments in the description text corresponding to the pictures in the second target picture cluster as text features or the feature vectors with the largest number of occurrences as text features. Furthermore, a more accurate degree of matching between the query terms and the text features corresponding to the pictures can be obtained, which is used to sort the search results, and finally a more accurate search result queue can be obtained.

It is understandable that, in addition to the image search request, the embodiment of the present application can also receive a webpage search request sent by the client and obtain the search results corresponding to the webpage search request. The information clustering result of the timeliness event is obtained by clustering the information generated within a preset time period in the network, and the information clustering result includes at least one text keyword. Then, according to the matching degree between the query term included in the search request and the text keyword, it is determined whether the webpage search request is a timeliness event search request.

When the search request is a web page search, the search results corresponding to the obtained web page search contain the keywords in the search request, and the query words included in the search request can be matched with the text keywords in the information clustering results, and the search request can be judged whether it is a time-sensitive event search request based on the matching degree between the query words and the text keywords. The method for determining the matching degree between the query words and the text keywords is not limited in the embodiments of the present application. In a possible implementation, the feature vector of the query words and the feature vector of the text keywords can be obtained, and the matching degree between the query words and the text keywords is determined by calculating the similarity between the feature vector of the query words and the feature vector of the text keywords.

Furthermore, the embodiment of the present application also provides a method for generating a search result queue by sorting the search results using the information clustering results when the webpage search request is a time-sensitive event search request. Referring to FIG. 5 , this figure is a flow chart of a method for generating a search result queue provided by an embodiment of the present application. As shown in FIG. 5 , the method may include S501-S502:

D1: When the webpage search request is a time-sensitive event search request, a search result including a text keyword in the search results corresponding to the webpage search request is obtained as a second target search result.

The search results corresponding to the web page search request are web page search results related to the query terms in the search request.

It is understandable that the text keyword is used to indicate relevant information of timeliness in the information clustering result. By using the search result with the text keyword as the second target search result, the search result with higher timeliness can be determined by the text keyword. In order to arrange the second target search result at a higher position in the search result queue during subsequent sorting, the user can quickly obtain the corresponding search result.

D2: Sort the second target search results and other search results in the search results according to the matching degree between the query term included in the webpage search request and the second target search results to generate a search result queue, wherein the ranking position of the second target search results is higher than the ranking positions of other search results in the search results.

The second target search result is a search result with text keywords and higher timeliness. The second target search result and other search results in the search results are sorted according to the degree of match between the second target search result and the query word. The second target search result with a higher degree of match with the query word can be placed at a higher sorting position. Thus, the search result with higher timeliness and a higher degree of relevance to the search request can be placed at a higher sorting position in the search result queue, so that the sorting of the search result queue is more accurate.

In the embodiment of the present application, when the search request is a webpage search request and a time-sensitive event search request, the second target search result is first determined by the text keywords, and then the second target search result and other search results in the search results are sorted according to the degree of matching between the query terms included in the search request and the second target search result. Thus, the search results that have a high degree of matching with the query terms included in the search request and are more timely can be arranged in a higher ranking position, and a search result queue with a more accurate ranking can be obtained, so that the user can quickly obtain the corresponding search results according to the more accurate search result queue.

Based on the search result sorting method provided by the above method embodiment, the embodiment of the present application also provides a search result sorting device, which is explained and illustrated below in conjunction with the accompanying drawings.

See Figure 5, which is a schematic diagram of the structure of a search result sorting device provided in an embodiment of the present application. The search result sorting device provided in an embodiment of the present application includes:

The search result acquisition unit 501 is used to receive an image search request sent by a client and acquire search results corresponding to the image search request;

An information clustering result acquisition unit 502 is used to acquire information clustering results of time-sensitive events, wherein the information clustering results of time-sensitive events are obtained by clustering information generated in a preset time period in the network, and the information clustering results include at least one picture cluster, and each of the picture clusters includes at least one picture;

A judging unit 503 is used to judge whether the image search request is a timeliness event search request according to the matching degree between the search result corresponding to the image search request and the image cluster;

The search result queue generating unit 504 is configured to, when the image search request is a time-sensitive event search request, sort the search results using the information clustering result to generate a search result queue.

In a possible implementation, the determining unit includes:

A first quantity determination module, configured to determine a first quantity of images in each of the image clusters that appear in the search results corresponding to the image search request;

A picture cluster determination module, configured to determine picture clusters whose first number is greater than a first threshold as first target picture clusters;

The first judgment module is configured to judge that the image search request is a timeliness event search request if the number of first target image clusters to which the search results corresponding to the image search request belong is greater than a second threshold.

In a possible implementation, the search result queue generating unit 504 includes:

A cluster feature acquisition subunit is used to acquire, when the image search request is a time-sensitive event search request, a cluster feature of a second target image cluster to which the search result corresponding to the image search request belongs; the second target image cluster is an image cluster corresponding to each image in the search result corresponding to the image search request;

a sorting result generating subunit, configured to sort the second target image clusters according to the cluster features of the second target image clusters to generate a sorting result;

A first selection subunit, configured to select a picture in each of the second target picture clusters as a first target search result;

The first search result queue generating subunit is used to sort the first target search result and other search results in the search results that do not belong to the second target image cluster according to the sorting result, and generate a search result queue, in which the sorting position of the first target search result is higher than the sorting position of the other search results.

In a possible implementation, the cluster feature acquisition subunit includes:

A source website quantity acquisition module, used to acquire a second quantity of images in a second target image cluster to which the search result corresponding to the image search request belongs, and a quantity of source websites corresponding to the images in the second target image cluster;

A text feature acquisition module, used to acquire text features corresponding to the pictures in the second target picture cluster;

The sorting result generating subunit is specifically used for:

The second target image cluster is sorted according to the matching degree between the query term included in the search request and the text features corresponding to the images in the second target image cluster, the second number of images in the second target image cluster, and the number of source websites corresponding to the images in the second target image cluster to generate a sorting result.

In a possible implementation, the text feature acquisition module is specifically used to:

Obtaining the word frequency of each word in the description text corresponding to the picture in the second target picture cluster, and taking at least one word with the highest word frequency as the text feature corresponding to the picture in the second target picture cluster;

Alternatively, the feature vectors of each word in the description text corresponding to the picture in the second target picture cluster are obtained, and at least one feature vector with the largest number of occurrences in the feature vectors is used as the text feature corresponding to the picture in the second target picture cluster.

In a possible implementation manner, the device further includes:

A spam information cluster determination unit, configured to determine a spam information cluster from the information clustering result of the timeliness event;

A spam information cluster removal unit, used to remove the spam information cluster from the information clustering result of the timeliness event;

The junk information cluster determination unit is specifically used to:

If a third number of pictures in the third target picture cluster is greater than a third threshold, or the number of source websites corresponding to the third target picture cluster is less than a fourth threshold, the third target picture cluster is determined as a spam information cluster, and the third target picture cluster is a picture cluster included in the information clustering result of the timeliness event;

Alternatively, if the similarities between the description texts corresponding to the pictures in the third target picture cluster are all less than the fifth threshold, the third target picture cluster is determined to be a spam cluster, and the third target picture cluster is the picture cluster included in the information clustering result of the timeliness event.

6 shows a block diagram of a search result ranking apparatus 1200. For example, the apparatus 1200 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

6 , the device 1200 may include one or more of the following components: a processing component 1202 , a memory 1204 , a power component 1206 , a multimedia component 1208 , an audio component 1210 , an input/output (I/O) interface 1212 , a sensor component 1214 , and a communication component 1216 .

The processing component 1202 generally controls the overall operation of the device 1200, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 1202 may include one or more processors 1220 to execute instructions to perform all or part of the steps of the above-described method. In addition, the processing component 1202 may include one or more modules to facilitate interaction between the processing component 1202 and other components. For example, the processing component 1202 may include a multimedia module to facilitate interaction between the multimedia component 1208 and the processing component 1202.

The memory 1204 is configured to store various types of data to support operations on the device 1200. Examples of such data include instructions for any application or method operating on the device 1200, contact data, phone book data, messages, pictures, videos, etc. The memory 1204 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

The power supply component 1206 provides power to the various components of the device 1200. The power supply component 1206 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device 1200.

The multimedia component 1208 includes a screen that provides an output interface between the device 1200 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 1208 includes a front camera and/or a rear camera. When the device 1200 is in an operating mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

The audio component 1210 is configured to output and/or input audio signals. For example, the audio component 1210 includes a microphone (MIC), and when the device 1200 is in an operation mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 1204 or sent via the communication component 1216. In some embodiments, the audio component 1210 also includes a speaker for outputting audio signals.

The I/O interface provides an interface between the processing component 1202 and the peripheral interface module, which may be a keyboard, a click wheel, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

The sensor assembly 1214 includes one or more sensors for providing various aspects of the status assessment of the device 1200. For example, the sensor assembly 1214 can detect the open/closed state of the device 1200, the relative positioning of components, such as the display and keypad of the device 1200, the sensor assembly 1214 can also detect the position change of the device 1200 or a component of the device 1200, the presence or absence of user contact with the device 1200, the orientation or acceleration/deceleration of the device 1200, and the temperature change of the device 1200. The sensor assembly 1214 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 1214 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 1214 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 1216 is configured to facilitate wired or wireless communication between the device 1200 and other devices. The device 1200 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 1216 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 1216 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

In an exemplary embodiment, the apparatus 1200 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the following methods:

Receive an image search request sent by a client, and obtain search results corresponding to the image search request;

Acquire information clustering results of time-sensitive events, where the information clustering results of time-sensitive events are obtained by clustering information generated in a preset time period in the network, and the information clustering results include at least one picture cluster, and each of the picture clusters includes at least one picture;

Determining whether the image search request is a timeliness event search request according to a matching degree between the search result corresponding to the image search request and the image cluster;

When the image search request is a time-sensitive event search request, the search results are sorted using the information clustering results to generate a search result queue.

Optionally, judging whether the image search request is a timeliness event search request according to a matching degree between the search result corresponding to the image search request and the image cluster includes:

Determine a first number of images in each of the image clusters that appear in the search results corresponding to the image search request;

Determine the picture clusters whose first number is greater than a first threshold as first target picture clusters;

If the number of first target image clusters to which the search results corresponding to the image search request belong is greater than a second threshold, it is determined that the image search request is a timeliness event search request.

Optionally, when the image search request is a time-sensitive event search request, sorting the search results using the information clustering results to generate a search result queue includes:

When the image search request is a time-sensitive event search request, obtaining cluster features of a second target image cluster to which the search results corresponding to the image search request belong; the second target image cluster is an image cluster corresponding to each image in the search results corresponding to the image search request;

sorting the second target image clusters according to the cluster features of the second target image clusters to generate a sorting result;

Selecting a picture in each of the second target picture clusters as the first target search result;

According to the sorting result, the first target search result and other search results in the search results that do not belong to the second target image cluster are sorted to generate a search result queue, in which the sorting position of the first target search result is higher than the sorting positions of the other search results.

Optionally, obtaining the cluster feature of the second target image cluster to which the search result corresponding to the image search request belongs includes:

Obtain a second number of images in a second target image cluster to which the search results corresponding to the image search request belong, and a number of source websites corresponding to the images in the second target image cluster;

Obtaining text features corresponding to the pictures in the second target picture cluster;

The step of sorting the target image clusters according to the cluster features of the second target image clusters to generate a sorting result includes:

The second target image cluster is sorted according to the matching degree between the query term included in the search request and the text features corresponding to the images in the second target image cluster, the second number of images in the second target image cluster, and the number of source websites corresponding to the images in the second target image cluster to generate a sorting result.

Optionally, obtaining text features corresponding to the pictures in the second target picture cluster includes:

Obtaining the word frequency of each word in the description text corresponding to the picture in the second target picture cluster, and taking at least one word with the highest word frequency as the text feature corresponding to the picture in the second target picture cluster;

Alternatively, the feature vectors of each word in the description text corresponding to the picture in the second target picture cluster are obtained, and at least one feature vector with the largest number of occurrences in the feature vectors is used as the text feature corresponding to the picture in the second target picture cluster.

Optionally, the method further includes:

Determining a junk information cluster in the information clustering result of the timeliness event, and removing the junk information cluster in the information clustering result of the timeliness event;

The step of determining a spam information cluster from the information clustering result of the timeliness event includes:

If a third number of pictures in the third target picture cluster is greater than a third threshold, or the number of source websites corresponding to the third target picture cluster is less than a fourth threshold, the third target picture cluster is determined as a spam information cluster, and the third target picture cluster is a picture cluster included in the information clustering result of the timeliness event;

Alternatively, if the similarities between the description texts corresponding to the pictures in the third target picture cluster are all less than the fifth threshold, the third target picture cluster is determined to be a spam cluster, and the third target picture cluster is the picture cluster included in the information clustering result of the timeliness event.

FIG7 is a schematic diagram of the structure of a server in an embodiment of the present invention. The server 1300 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 1322 (for example, one or more processors) and memory 1332, and one or more storage media 1330 (for example, one or more mass storage devices) storing application programs 1342 or data 1344. Among them, the memory 1332 and the storage medium 1330 may be temporary storage or permanent storage. The program stored in the storage medium 1330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processing unit 1322 may be configured to communicate with the storage medium 1330, and execute a series of instruction operations in the storage medium 1330 on the server 1300 for executing the above-mentioned search result sorting method.

The server 1300 may also include one or more power supplies 1326, one or more wired or wireless network interfaces 1350, one or more input and output interfaces 1356, one or more keyboards 1356, and/or, one or more operating systems 1341, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

In addition, an embodiment of the present application also provides a computer-readable medium having instructions stored thereon, which, when executed by one or more processors, enables the device to perform the above-mentioned search result sorting method.

It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the system or device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And/or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and/or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character "/" generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims (12)

1.A method of ranking search results, the method comprising:

receiving a picture searching request sent by a client, and acquiring a searching result corresponding to the picture searching request;

Acquiring an information clustering result of a timeliness event, wherein the information clustering result of the timeliness event is obtained by clustering information generated in a preset time period in a network, the information clustering result comprises at least one picture cluster, and each picture cluster comprises at least one picture;

Determining a first number of pictures in each picture class cluster appearing in search results corresponding to the picture search request;

determining the first number of picture class clusters larger than a first threshold as a first target picture class cluster;

If the number of the first target picture clusters to which the search results corresponding to the picture search requests belong is larger than a second threshold, judging that the picture search requests are timeliness event search requests;

When the picture searching request is a time-efficient event searching request, sequencing the searching results by utilizing the information clustering results to generate a searching result queue; in the search result queue, the higher the search result ranking is, the higher the association degree between the search result and the timeliness event is, and the higher the timeliness is.

2. The method of claim 1, wherein when the picture search request is a time-lapse event search request, ranking the search results using the information cluster results to generate a search result queue, comprising:

When the picture searching request is a time-based event searching request, acquiring class cluster characteristics of a second target picture class cluster to which a searching result corresponding to the picture searching request belongs; the second target picture cluster is a picture cluster corresponding to each picture in the search result corresponding to the picture search request;

sorting the second target picture clusters according to the cluster characteristics of the second target picture clusters to generate a sorting result;

selecting one picture from each second target picture cluster as a first target search result;

And according to the sorting results, sorting the first target search results and other search results which do not belong to the second target picture cluster in the search results, and generating a search result queue, wherein the sorting position of the first target search results is higher than the sorting position of the other search results in the search result queue.

3. The method according to claim 2, wherein the obtaining the cluster feature of the second target picture cluster to which the search result corresponding to the picture search request belongs includes:

acquiring a second number of pictures in a second target picture cluster to which a search result corresponding to the picture search request belongs and a source website number corresponding to the pictures in the second target picture cluster;

Acquiring text features corresponding to pictures in the second target picture cluster;

The step of sorting the target picture clusters according to the cluster characteristics of the second target picture clusters to generate a sorting result comprises the following steps:

And sorting the second target picture clusters according to the matching degree of the query words included in the search request and the text features corresponding to the pictures in the second target picture clusters, the second number of the pictures in the second target picture clusters and the number of source websites corresponding to the pictures in the second target picture clusters, and generating a sorting result.

4. The method according to claim 3, wherein the obtaining text features corresponding to the pictures in the second target picture class cluster includes:

acquiring word frequency of each word segmentation in the descriptive text corresponding to the picture in the second target picture cluster, and taking at least one word segmentation with the highest word frequency as a text feature corresponding to the picture in the second target picture cluster;

Or obtaining feature vectors of each word in the descriptive text corresponding to the pictures in the second target picture cluster, wherein at least one feature vector with the largest occurrence frequency in the feature vectors is used as the text feature corresponding to the pictures in the second target picture cluster.

5. The method according to any one of claims 1-4, further comprising:

Determining a junk information cluster in the information clustering result of the time-efficient event, and removing the junk information cluster in the information clustering result of the time-efficient event;

The determining the junk information cluster in the information clustering result of the time-lapse event comprises the following steps:

If the third number of pictures in the third target picture cluster is larger than a third threshold value or the number of source websites corresponding to the third target picture cluster is smaller than a fourth threshold value, determining the third target picture cluster as a junk information cluster, wherein the third target picture cluster is a picture cluster included in an information clustering result of the timeliness event;

or if the similarity between the descriptive texts corresponding to the pictures in the third target picture cluster is smaller than a fifth threshold, determining the third target picture cluster as a junk information cluster, wherein the third target picture cluster is a picture cluster included in the information clustering result of the timeliness event.

6. A search result ordering apparatus, the apparatus comprising:

The device comprises a search result acquisition unit, a search result generation unit and a search result generation unit, wherein the search result acquisition unit is used for receiving a picture search request sent by a client and acquiring a search result corresponding to the picture search request;

The information clustering result acquisition unit is used for acquiring an information clustering result of the timeliness event, wherein the information clustering result of the timeliness event is obtained by clustering information generated in a preset time period in a network, the information clustering result comprises at least one picture cluster, and each picture cluster comprises at least one picture;

The judging unit is used for judging whether the picture searching request is an aging event searching request or not according to the matching degree of the searching result corresponding to the picture searching request and the picture cluster;

The search result queue generating unit is used for sequencing the search results by utilizing the information clustering results when the picture search request is a time-lapse event search request, so as to generate a search result queue, wherein the higher the search result is sequenced, the higher the association degree between the search result and the time-lapse event is, and the higher the time-lapse event is;

The judging unit includes:

A first number determining module, configured to determine a first number of pictures in each of the picture clusters appearing in a search result corresponding to the picture search request;

the picture cluster determining module is used for determining the picture clusters with the first number larger than a first threshold value as first target picture clusters;

the first judging module is configured to judge that the picture search request is an aging event search request if the number of first target picture clusters to which the search result corresponding to the picture search request belongs is greater than a second threshold.

7. The apparatus of claim 6, wherein the search result queue generating unit comprises:

a cluster feature obtaining subunit, configured to obtain a cluster feature of a second target picture cluster to which a search result corresponding to the picture search request belongs when the picture search request is an aging event search request; the second target picture cluster is a picture cluster corresponding to each picture in the search result corresponding to the picture search request;

The sequencing result generation subunit is used for sequencing the second target picture cluster according to the cluster characteristics of the second target picture cluster to generate a sequencing result;

the first selecting subunit is used for selecting one picture from each second target picture cluster as a first target search result;

And the first search result queue generating subunit is used for sequencing the first target search result and other search results which do not belong to the second target picture cluster in the search results according to the sequencing results, so as to generate a search result queue, wherein the sequencing position of the first target search result in the search result queue is higher than the sequencing position of the other search results.

8. The apparatus of claim 7, wherein the cluster-like feature acquisition subunit comprises:

The source website number acquisition module is used for acquiring the second number of pictures in a second target picture cluster to which the search result corresponding to the picture search request belongs and the source website number corresponding to the pictures in the second target picture cluster;

the text feature acquisition module is used for acquiring text features corresponding to the pictures in the second target picture cluster;

The sequencing result generation subunit is specifically configured to:

And sorting the second target picture clusters according to the matching degree of the query words included in the search request and the text features corresponding to the pictures in the second target picture clusters, the second number of the pictures in the second target picture clusters and the number of source websites corresponding to the pictures in the second target picture clusters, and generating a sorting result.

9. The apparatus of claim 8, wherein the text feature acquisition module is specifically configured to:

acquiring word frequency of each word segmentation in the descriptive text corresponding to the picture in the second target picture cluster, and taking at least one word segmentation with the highest word frequency as a text feature corresponding to the picture in the second target picture cluster;

Or obtaining feature vectors of each word in the descriptive text corresponding to the pictures in the second target picture cluster, wherein at least one feature vector with the largest occurrence frequency in the feature vectors is used as the text feature corresponding to the pictures in the second target picture cluster.

10. The apparatus according to any one of claims 6-9, wherein the apparatus further comprises:

The junk information cluster determining unit is used for determining junk information clusters in the information clustering result of the time-lapse event;

The junk information cluster removing unit is used for removing the junk information clusters from the information clustering result of the timeliness event;

the garbage information cluster determining unit is specifically configured to:

If the third number of pictures in the third target picture cluster is larger than a third threshold value or the number of source websites corresponding to the third target picture cluster is smaller than a fourth threshold value, determining the third target picture cluster as a junk information cluster, wherein the third target picture cluster is a picture cluster included in an information clustering result of the timeliness event;

or if the similarity between the descriptive texts corresponding to the pictures in the third target picture cluster is smaller than a fifth threshold, determining the third target picture cluster as a junk information cluster, wherein the third target picture cluster is a picture cluster included in the information clustering result of the timeliness event.

11. A search result ordering apparatus comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising instructions for:

receiving a picture searching request sent by a client, and acquiring a searching result corresponding to the picture searching request;

Acquiring an information clustering result of a timeliness event, wherein the information clustering result of the timeliness event is obtained by clustering information generated in a preset time period in a network, the information clustering result comprises at least one picture cluster, and each picture cluster comprises at least one picture;

Determining a first number of pictures in each picture class cluster appearing in search results corresponding to the picture search request;

determining the first number of picture class clusters larger than a first threshold as a first target picture class cluster;

If the number of the first target picture clusters to which the search results corresponding to the picture search requests belong is larger than a second threshold, judging that the picture search requests are timeliness event search requests;

When the picture searching request is a time-efficient event searching request, sequencing the searching results by utilizing the information clustering results to generate a searching result queue; in the search result queue, the higher the search result ranking is, the higher the association degree between the search result and the timeliness event is, and the higher the timeliness is.

12. A computer-readable medium having instructions stored thereon, which when executed by one or more processors, cause an apparatus to perform the search result ordering method of one or more of claims 1 to 5.