CN112307337B

CN112307337B - Associated recommendation method and device based on tag knowledge graph and computer equipment

Info

Publication number: CN112307337B
Application number: CN202011193407.2A
Authority: CN
Inventors: 张松
Original assignee: Ping An Life Insurance Company of China Ltd
Current assignee: Ping An Life Insurance Company of China Ltd
Priority date: 2020-10-30
Filing date: 2020-10-30
Publication date: 2024-04-12
Anticipated expiration: 2040-10-30
Also published as: CN112307337A

Abstract

The invention discloses a label knowledge graph-based associated recommendation method, a device and computer equipment, wherein the method comprises the following steps: acquiring a text type of each text in a text set; marking each text according to the text type of each text to obtain a plurality of basic labels marked on each text; inputting each text marked with a plurality of basic labels into a deep learning model to obtain a high-order semantic label of each text; constructing a tag knowledge graph of a text set according to the high-order semantic tags, attribute information of the high-order semantic tags and relation information among the high-order semantic tags; if the input information of the user is received, a high-order semantic tag matched with the input information is obtained from the tag knowledge graph according to the input information so as to carry out information pushing service on the user. The invention belongs to the technical field of artificial intelligence, and can accurately push according to the input information of a user, thereby improving the pushing accuracy.

Description

Associated recommendation method and device based on tag knowledge graph and computer equipment

Technical Field

The invention belongs to the technical field of artificial intelligence, and particularly relates to a tag knowledge graph-based associated recommendation method, a tag knowledge graph-based associated recommendation device and computer equipment.

Background

The knowledge map is essentially a semantic network for revealing the relation between entities, called a knowledge domain visualization or knowledge domain mapping map in book information, is a series of different graphs for displaying the knowledge development process and the structural relation, describes knowledge resources and carriers thereof by using a visualization technology, and excavates, analyzes, builds, draws and displays knowledge and the relation among the knowledge resources and the carriers. A user usually uses a knowledge graph to acquire knowledge, and the knowledge graph in the prior art has the problems of sparsity and cold start in a specific field. Sparsity refers to the fact that when the interactive information of a user and an article is sparse, less observation data are used for predicting a large amount of unknown information, and the risk of overfitting is greatly increased; cold start refers to the difficulty in using knowledge maps for accurate modeling and recommendation for newly added users or items when there is no corresponding historical information.

Disclosure of Invention

The embodiment of the invention provides a method, a device and computer equipment for associated recommendation based on a tag knowledge graph, which solve the problem that the prior art cannot accurately recommend users.

In a first aspect, an embodiment of the present invention provides a method for associated recommendation based on a tag knowledge graph, including:

content analysis is carried out on each text in a preset text set according to a preset analysis model so as to obtain the text type of each text;

marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels;

inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text;

constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags;

and if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user.

In a second aspect, an embodiment of the present invention provides an association recommendation device based on a tag knowledge graph, including:

the analysis unit is used for carrying out content analysis on each text in a preset text set according to a preset analysis model so as to obtain the text type of each text;

The labeling unit is used for labeling each text according to the text type of each text so as to obtain a text labeled with a plurality of basic labels;

the first input unit is used for inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text;

the construction unit is used for constructing a tag knowledge graph of the text set according to the higher-order semantic tags of each text, the attribute information of the higher-order semantic tags and the relation information among the higher-order semantic tags;

and the recommending unit is used for acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information if the input information of the user is received, so as to carry out information pushing service on the user.

In a third aspect, an embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor implements the association recommendation method based on the tag knowledge pattern according to the first aspect when executing the computer program.

In a fourth aspect, an embodiment of the present invention further provides a computer readable storage medium, where the computer readable storage medium stores a computer program, where the computer program when executed by a processor causes the processor to perform the association recommendation method based on the tag knowledge pattern according to the first aspect.

The embodiment of the invention provides a method, a device and computer equipment for associated recommendation based on a tag knowledge graph, which are used for carrying out content analysis on each text in a preset text set through a preset analysis model so as to obtain the text type of each text; marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels; inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text; constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags; and if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user. According to the associated recommendation method based on the tag knowledge graph, the tag knowledge graph of the high-order semantic meaning is constructed, so that after the input information of the user is obtained, the information required by the user can be accurately obtained from the tag knowledge graph according to the input information of the user and pushed to the user, and the pushing accuracy is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flow chart of a related recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

fig. 2 is a schematic view of a scenario of an association recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

fig. 3 is a schematic sub-flowchart of an association recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

fig. 4 is another schematic sub-flowchart of a related recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

fig. 5 is another schematic sub-flowchart of an association recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

fig. 6 is another schematic sub-flowchart of a related recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

fig. 7 is another schematic sub-flowchart of a related recommendation method based on a tag knowledge graph according to an embodiment of the present invention;

FIG. 8 is a schematic block diagram of an associated recommendation device based on a tag knowledge graph according to an embodiment of the present invention;

FIG. 9 is a schematic block diagram of a subunit of a tag knowledge-graph-based associative recommendation device according to an embodiment of the present invention;

FIG. 10 is a schematic block diagram of another subunit of the associated recommendation device based on a tag knowledge-graph according to an embodiment of the present invention;

FIG. 11 is a schematic block diagram of another subunit of the associated recommendation device based on a tag knowledge pattern according to an embodiment of the present invention;

FIG. 12 is a schematic block diagram of another subunit of the associated recommendation device based on a tag knowledge-graph according to an embodiment of the present invention;

FIG. 13 is a schematic block diagram of another subunit of the associated recommendation device based on a tag knowledge-graph according to an embodiment of the present invention;

fig. 14 is a schematic block diagram of a computer device according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.

Referring to fig. 1 and fig. 2, fig. 1 is a flow chart of a related recommendation method based on a tag knowledge graph according to an embodiment of the present invention; fig. 2 is a schematic view of a scenario of an association recommendation method based on a tag knowledge graph according to an embodiment of the present invention. The associated recommendation method based on the tag knowledge graph is built and operated in a server, and the server builds the tag knowledge graph through a preset text set. After receiving input information sent by a user through user terminal equipment such as a portable computer, a tablet personal computer and the like, the server analyzes the input information to obtain a high-order semantic tag of the input information and obtains related information of the high-order semantic tag through a tag knowledge graph so as to carry out accurate push service on the user.

As shown in fig. 1, the method includes steps S110 to S150.

S110, carrying out content analysis on each piece of text in a preset text set according to a preset analysis model to obtain the text type of each piece of text.

And carrying out content analysis on each text in the preset text set according to the preset analysis model to obtain the text type of each text. Specifically, the parsing model is a model for performing content parsing on each text in the text set to obtain a text type of each text. The text type is any one of a material type, an objection type and a FAQ type. The text set includes at least one story type of text, at least one objection type of text, at least one FAQ type of text. The text of the material type is the text of the material type required by the target scheme, the text of the objection type is the text of the user questioning and answering type, and the text of the FAQ type is the text of the user consultation and consultation answering type. The target scheme at least comprises a text of a material type, the question and the solution of the question corresponding to the target scheme at least comprise a text of an objection type, and the consultation and the solution of the consultation corresponding to the target scheme at least comprise a text of a FAQ type.

For example, in the security industry, when an insurance agent needs to recommend an dangerous seed to a customer, the customer agent needs to make a target scheme about the dangerous seed, where the target scheme of the dangerous seed includes at least one text of a material type related to the dangerous seed, and when the insurance agent explains to the customer according to the target scheme of the dangerous seed, if the customer makes a question according to the target scheme of the dangerous seed, the insurance agent makes a solution according to the question, and the question of the customer and the solution of the insurance agent to the question include at least one text of an objection type; if the client makes a consultation according to the dangerous target scheme, the insurance agent makes a solution according to the consultation, and the consultation of the client and the solution of the insurance agent to the consultation at least comprise a text of FAQ type.

And S120, marking each text according to the text type of each text to obtain texts marked with a plurality of basic labels.

And marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels. Specifically, each text in the text set is labeled with at least one base label. After each text in the text set and the text type of each text are obtained, the text marking tool marks the text according to the text type of each text so that each text in the data set is marked with a plurality of basic labels matched with the text type of the text. When the text type of one text in the text set is the text of the material type, the text labeling tool labels the text so that the text is labeled with a plurality of basic labels matched with the material type, wherein the plurality of basic labels matched with the material type comprise basic labels such as gender, annual income, product highlight, commodity key words, activity type, service type, task value, exhibition industry habit, sales stage, guarantee range, content type, activity range, tool type, game type, user portrait, agent portrait, mechanism portrait and the like; when the text type of one text in the text set is a text of an objection type, the text marking tool marks the text so that the text is marked with a plurality of basic labels matched with the objection type, wherein the basic labels matched with the objection type comprise basic labels such as gender, annual income, product highlight, commodity key words, activity type, service type, task value, exhibition industry habit, sales stage, guarantee range, content type, activity range, tool type, game type, user portrait, agent portrait, mechanism portrait and the like; when the text type of one text in the text set is the text of the FAQ type, the text labeling tool labels the text so that the text is labeled with a plurality of basic labels matched with the material type, wherein the basic labels matched with the FAQ type comprise basic labels of gender, annual income, product highlight, commodity keywords, activity type, service type, task value, exhibition industry habit, sales stage, guarantee range, content type, activity range, tool type, game type, user portrait, agent portrait, mechanism portrait and the like.

S130, inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text.

And inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text. Specifically, the deep learning model is a model for acquiring semantic information of each text from each text marked with a plurality of basic labels and extracting high-order semantic labels of each text from the plurality of basic labels in each text according to the semantic information. The high-order semantic tags are obtained by associating semantic information of each text with a plurality of basic tags corresponding to the text. For example, a text of a material type is "a woman who has born a woman who is two tens of thousands of people in the month", a plurality of basic labels of the text are respectively "month income", "marital", "gender", "birth", and the basic labels of the text do not have direct relation, and the basic labels of the text are associated by semantic information of the text, so that a high-order semantic label of "job-site-treasure mom" is extracted from the basic labels of the text.

In one embodiment, as shown in FIG. 3, step S130 includes sub-steps S131, S132, and S133.

S131, word segmentation processing is carried out on each text marked with a plurality of basic labels to obtain words of each text.

And performing word segmentation processing on each text marked with a plurality of basic labels to obtain words of each text. Specifically, a word segmentation process is carried out on each text marked with a plurality of basic labels by adopting a reverse maximum matching method in a word segmentation method based on character strings, and the word segmentation process comprises the following steps: the number of Chinese characters contained in the longest entry in the preset dictionary is set as L, and the processing is started from the end of the character string of each text. And when each cycle starts, the last L words of the character string are taken as processing objects, and the dictionary is searched. If the dictionary contains such an L word, the matching is successful, and the processing object is segmented as a word; if not, the first Chinese character of the processing object is removed, the rest character strings are used as new processing objects, matching is carried out again until segmentation is successful, namely, one round of matching is completed, one word is segmented, and the cycle is like until all words in the text are segmented.

For example, the length of the longest word in the dictionary is 6, and for a character string with the text of 'female with childbearing' at home, '6 characters with the character at home' are taken as the character string to be processed, and the word is not in the dictionary, so that the matching is failed; removing the first word, taking the rest 'at home' as a new character string to be processed, and failing to match again; and finally, going to 'to-be-employment' as a matching field, wherein if the word exists in the dictionary, the matching is successful, and the first word at the segmentation position is 'to-be-employment'. The remaining string "female born at home" in the text is then taken, and the second word "at home" is cut. The cycle is thus completed, and the final segmentation results are: "fertile", "female", "at home", "to-be-employment".

S132, inputting the words of each text into a pre-trained language model to obtain semantic information of each text.

And inputting the words of each text into a pre-trained language model to obtain semantic information of each text. Specifically, the language model is a model which is trained in advance and is used for carrying out semantic analysis on the words of each text so as to obtain semantic information of each text. And analyzing and identifying the words of each text and the position relation among the words through the language model to obtain the semantic information of each text. The language model may be any one of BERT (Bidirectional Encoder Representations from Transformers) model and ELMO (Embedding from language model) model.

S133, associating a plurality of basic labels in each text according to the semantic information of each text to obtain a high-order semantic label of each text.

And associating a plurality of basic labels in each text according to the semantic information of each text to obtain the high-order semantic label of each text. Specifically, there is generally no direct association between the plurality of basic labels in each text, and an indirect relationship of the plurality of basic labels in the text is abstracted from the plurality of basic labels in the text through the semantic information of each text, where the indirect relationship is the high-order semantic label of the text.

S140, constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags.

And constructing a tag knowledge graph of the text set according to the higher-order semantic tags of each text, the attribute information of the higher-order semantic tags and the relation information among the higher-order semantic tags. Specifically, the tag knowledge graph is a knowledge graph constructed by taking a higher-order semantic tag of each text in the text set as an entity, and taking a relationship between the higher-order semantic tag and the higher-order semantic tag of each text in the text set as a relationship between the entities. The tag knowledge graph in the embodiment of the invention is essentially a semantic network for revealing the relationship between the higher-order semantic tags and the higher-order semantic tags of each text in the text set.

In one embodiment, as shown in FIG. 4, step S140 includes sub-steps S141 and S142.

S141, performing knowledge grabbing and entity linking on the higher-order semantic tags of each text, the attribute information of the higher-order semantic tags and the relation information among the higher-order semantic tags to obtain the tag knowledge graph information.

And carrying out knowledge grabbing and entity linking on the higher-order semantic tags of each text, the attribute information of the higher-order semantic tags and the relation information among the higher-order semantic tags to obtain the tag knowledge graph information. Specifically, the higher-order semantic tags, the attribute of the higher-order semantic tags and the association relation among the higher-order semantic tags are captured from the higher-order semantic tags, the attribute information of the higher-order semantic tags and the relation information among the higher-order semantic tags, then the captured knowledge is subjected to higher-order semantic tag connection, namely the knowledge is integrated, and tag knowledge map information with relation among the higher-order semantic tags is formed.

S142, storing the tag knowledge graph information into a structured knowledge base to form the tag knowledge graph.

And storing the tag knowledge-graph information into a structured knowledge base to form the tag knowledge-graph. Specifically, the tag knowledge graph information obtained after integration is added into a structured knowledge base, knowledge graph demonstration is carried out by utilizing knowledge reasoning of the knowledge base, and thus the tag knowledge graph information is obtained.

And S150, if input information of a user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user.

And if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user. Specifically, the input information is analyzed to obtain a plurality of basic labels and semantic information of the input information, and the basic labels in the basic labels of the input information are associated according to the semantic information of the input information so as to obtain a high-order semantic label matched with the input information from the label knowledge graph, so that text information matched with the input information is obtained from the text set.

In one embodiment, as shown in FIG. 5, step S150 includes sub-steps S151 and S152.

S151, judging the type of the input information.

And judging the type of the input information. Specifically, the input information of the user may be text information or voice information, where the voice information is a word that needs to be converted into text information and then processed to obtain the input information in the subsequent step, so that the terminal needs to determine the input information of the user after obtaining the input information of the user to obtain the type of the input information.

S152, if the type of the input information is voice input information, converting the voice input information into text information of the input information, and acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the text information.

If the type of the input information is voice input information, converting the voice input information into text information of the input information and acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the text information. Specifically, when a user inputs information from a voice information collection function at the user terminal, the user terminal can determine that the input information is voice input information, and then can convert the voice input information to obtain text information corresponding to the voice information, namely, the text information of the input information.

In one embodiment, as shown in fig. 6, the obtaining, in step S152, the high-order semantic tags matched with the input information from the tag knowledge graph according to the text information includes: steps S1521, S1522, and S1523.

S1521, acquiring semantic information of the text information and a plurality of basic labels of the text information.

And acquiring semantic information of the text information and a plurality of basic labels of the text information. Specifically, the text information is analyzed to obtain the text type of the text information, then the text information is marked according to the text type of the text information to obtain a plurality of basic labels of the text information, and the marked text information is subjected to semantic analysis by using a language model to obtain semantic information of the text information.

S1522, associating a plurality of basic labels of the text information according to the semantic information of the text information to obtain the advanced semantic labels of the input information.

In one embodiment, as shown in fig. 7, step S1522 includes sub-steps S15221, S15222, and S15223.

S15221, judging whether the input information is the input information input for the first time.

And judging whether the input information is the input information input for the first time. Specifically, whether the input information is the first input information is discriminated by the text type of the input information. When the input information is any one of an objection type and a FAQ type, judging that the input information is not the input information input for the first time; when the input information is the information of the material type, the input information can be judged to be the input information input for the first time.

S15222, if the input information is not the input information input for the first time, acquiring a plurality of basic labels before the input information according to a preset acquisition rule.

If the input information is not the input information input for the first time, acquiring a plurality of basic labels before the input information according to a preset acquisition rule. Specifically, the preset acquiring rule is rule information of a plurality of basic tags inputting information at the terminal within a preset time before acquiring the input information. The semantic information of the current input information is used for associating the plurality of basic labels of the current input information with the plurality of basic labels of the previous input information, and then the higher-level semantic labels are extracted, so that corresponding pushing service can be performed on the user more accurately.

S15223, associating the basic labels of the text information and the basic labels of the last moment according to the semantic information of the text information to obtain the advanced semantic label of the input information.

And associating the basic labels of the text information and the basic labels at the previous moment according to the semantic information of the text information to obtain the advanced semantic label of the input information. Specifically, the semantic information of the current input information can be used for extracting the high-order semantic tag of the current input information from a plurality of basic tags of the current input information and a plurality of basic tags of the previous input information.

S1523, obtaining the high-order semantic tags matched with the input information from the tag knowledge graph according to the high-order semantic tags of the input information.

According to the associated recommendation method based on the tag knowledge graph, content analysis is carried out on each text in a preset text set through a preset analysis model so as to obtain the text type of each text; marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels; inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text; constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags; and if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user. According to the associated recommendation method based on the tag knowledge graph, the tag knowledge graph of the high-order semantic meaning is constructed, so that after the input information of the user is obtained, the information required by the user can be accurately obtained from the tag knowledge graph according to the input information of the user and pushed to the user, and the pushing accuracy is improved.

The embodiment of the invention also provides a related recommendation device 100 based on the tag knowledge graph, which is used for executing any embodiment of the related recommendation method based on the tag knowledge graph. Specifically, referring to fig. 8, fig. 8 is a schematic block diagram of an association recommendation device 100 based on a tag knowledge graph according to an embodiment of the present invention.

As shown in fig. 8, the related recommending device 100 based on the tag knowledge graph includes an analyzing unit 110, a labeling unit 120, a first input unit 130, a constructing unit 140 and a recommending unit 150.

And the parsing unit 110 is configured to parse the content of each text in the preset text set according to the preset parsing model to obtain the text type of each text.

And the labeling unit 120 is configured to label each text according to the text type of each text, so as to obtain a text labeled with a plurality of basic labels.

A first input unit 130, configured to input each text labeled with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text.

In other inventive embodiments, as shown in fig. 9, the first input unit 130 includes: a word segmentation unit 131, a second input unit 132, and a first acquisition unit 133.

And the word segmentation unit 131 is configured to perform word segmentation processing on each text labeled with a plurality of basic labels to obtain words of each text.

And a second input unit 132, configured to input the word of each text into a pre-trained language model to obtain semantic information of each text.

The first obtaining unit 133 is configured to associate a plurality of basic labels in each text according to the semantic information of each text to obtain a high-order semantic label of each text.

And a construction unit 140, configured to construct a tag knowledge graph of the text set according to the higher-order semantic tags of each text, attribute information of the higher-order semantic tags, and relationship information between the higher-order semantic tags.

In other inventive embodiments, as shown in fig. 10, the construction unit 140 includes: a gripping unit 141 and a storage unit 142.

And a capturing unit 141, configured to perform knowledge capturing and entity linking on the higher-order semantic tags of each text, attribute information of the higher-order semantic tags, and relationship information between the higher-order semantic tags, so as to obtain the tag knowledge graph information.

And a storage unit 142, configured to store the tag knowledge-graph information into a structured knowledge base to form the tag knowledge-graph.

And the recommending unit 150 is configured to obtain, if input information of a user is received, a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information, so as to perform information pushing service on the user.

In other embodiments of the invention, as shown in fig. 11, the recommendation unit 150 includes: a first judgment unit 151, a conversion unit 152, and a second acquisition unit 153.

A first judging unit 151 for judging the type of the input information.

The converting unit 152 is configured to convert the voice input information into text information of the input information if the type of the input information is voice input information, and obtain, according to the text information, a high-order semantic tag matched with the input information from the tag knowledge graph.

In other inventive embodiments, as shown in fig. 12, the conversion unit 152 includes: a second acquisition unit 1521, a third acquisition unit 1522, and a fourth acquisition unit 1523.

A second obtaining unit 1521 is configured to obtain semantic information of the text information and a plurality of basic tags of the text information.

A third obtaining unit 1522, configured to associate a plurality of basic tags of the text information according to semantic information of the text information to obtain advanced semantic tags of the input information.

In other inventive embodiments, as shown in fig. 13, the third obtaining unit 1522 includes: a second determination unit 15221, a fifth acquisition unit 15222, and a sixth acquisition unit 15223.

The second judging unit 15221 is configured to judge whether the input information is the first input information.

The fifth obtaining unit 15222 is configured to obtain, according to a preset obtaining rule, a plurality of basic tags before the input information if the input information is not the input information input for the first time.

A sixth obtaining unit 15223, configured to associate, according to semantic information of the text information, a plurality of basic tags of the text information and the plurality of basic tags of the previous time to obtain advanced semantic tags of the input information.

A fourth obtaining unit 1523, configured to obtain, from the tag knowledge graph, a high-order semantic tag that matches the input information according to the high-order semantic tag of the input information.

The associated recommending device 100 based on the tag knowledge graph provided by the embodiment of the present invention is configured to perform the content parsing on each text in a preset text set through a preset parsing model to obtain a text type of each text; marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels; inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text; constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags; and if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user.

Referring to fig. 14, fig. 14 is a schematic block diagram of a computer device according to an embodiment of the present invention.

Referring to fig. 14, the device 500 includes a processor 502, a memory, and a network interface 505, which are connected by a system bus 501, wherein the memory may include a non-volatile storage medium 503 and an internal memory 504.

The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032, when executed, may cause the processor 502 to perform an associated recommendation method based on a tag knowledge graph.

The processor 502 is used to provide computing and control capabilities to support the operation of the overall device 500.

The internal memory 504 provides an environment for the execution of a computer program 5032 in the non-volatile storage medium 503, which computer program 5032, when executed by the processor 502, may cause the processor 502 to perform an associated recommendation method based on a tag knowledge graph.

The network interface 505 is used for network communication, such as providing for transmission of data information, etc. It will be appreciated by those skilled in the art that the structure shown in fig. 14 is merely a block diagram of a portion of the structure associated with the present inventive arrangements and is not limiting of the apparatus 500 to which the present inventive arrangements are applied, and that a particular apparatus 500 may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.

Wherein the processor 502 is configured to execute a computer program 5032 stored in a memory to perform the following functions: content analysis is carried out on each text in a preset text set through a preset analysis model so as to obtain the text type of each text; marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels; inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text; constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags; and if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user.

Those skilled in the art will appreciate that the embodiment of the apparatus 500 shown in fig. 14 is not limiting of the specific construction of the apparatus 500, and in other embodiments, the apparatus 500 may include more or less components than illustrated, or certain components may be combined, or a different arrangement of components. For example, in some embodiments, the device 500 may include only the memory and the processor 502, and in such embodiments, the structure and the function of the memory and the processor 502 are consistent with the embodiment shown in fig. 14, and will not be described herein.

It should be appreciated that in an embodiment of the invention, the processor 502 may be a central processing unit (Central Processing Unit, CPU), the processor 502 may also be other general purpose processors 502, digital signal processors 502 (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. Wherein the general purpose processor 502 may be the microprocessor 502 or the processor 502 may be any conventional processor 502 or the like.

In another embodiment of the invention, a computer storage medium is provided. The storage medium may be a non-volatile computer readable storage medium. The storage medium stores a computer program 5032, wherein the computer program 5032 when executed by the processor 502 performs the steps of: content analysis is carried out on each text in a preset text set through a preset analysis model so as to obtain the text type of each text; marking each text according to the text type of each text to obtain a text marked with a plurality of basic labels; inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text; constructing a tag knowledge graph of the text set according to the high-order semantic tags of each text, the attribute information of the high-order semantic tags and the relation information among the high-order semantic tags; and if the input information of the user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus, device and unit described above may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein. Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus, device and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, and for example, the division of the units is merely a logical function division, there may be another division manner in actual implementation, or units having the same function may be integrated into one unit, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices, or elements, or may be an electrical, mechanical, or other form of connection.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment of the present invention.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units may be stored in a storage medium if implemented in the form of software functional units and sold or used as stand-alone products. Based on such understanding, the technical solution of the present invention may be essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing an apparatus 500 (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, an optical disk, or other various media capable of storing program codes.

While the invention has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims

1. The associated recommendation method based on the tag knowledge graph is characterized by comprising the following steps of:

If input information of a user is received, acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information so as to perform information pushing service on the user;

the step of inputting each text marked with a plurality of basic labels into a preset deep learning model to obtain a high-order semantic label of each text comprises the following steps:

word segmentation processing is carried out on each text marked with a plurality of basic labels so as to obtain words of each text;

inputting the words of each text into a pre-trained language model to obtain semantic information of each text;

associating a plurality of basic labels in each text according to the semantic information of each text to obtain a high-order semantic label of each text;

the obtaining, according to the input information, a high-order semantic tag matched with the input information from the tag knowledge graph includes:

judging the type of the input information;

if the type of the input information is voice input information, converting the voice input information into text information of the input information and acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the text information;

The obtaining, according to the text information, a high-order semantic tag matched with the input information from the tag knowledge graph includes:

acquiring semantic information of the text information and a plurality of basic labels of the text information;

associating a plurality of basic tags of the text information according to the semantic information of the text information to obtain an advanced semantic tag of the input information;

and acquiring the high-order semantic tag matched with the input information from the tag knowledge graph according to the high-order semantic tag of the input information.

2. The tag knowledge graph-based associative recommendation method of claim 1, wherein the word segmentation process is performed on each text labeled with a plurality of basic tags to obtain words of each text, and the method comprises:

and performing word segmentation processing on each text marked with a plurality of basic labels according to a preset dictionary to obtain words of each text.

3. The tag knowledge graph-based associative recommendation method according to claim 1, wherein the constructing the tag knowledge graph of the text set according to the higher-order semantic tags of each text, attribute information of the higher-order semantic tags, and relationship information between the higher-order semantic tags includes:

Performing knowledge grabbing and entity linking on the higher-order semantic tags of each text, attribute information of the higher-order semantic tags and relation information among the higher-order semantic tags to obtain tag knowledge graph information;

and storing the tag knowledge-graph information into a structured knowledge base to form the tag knowledge-graph.

4. The tag knowledge graph-based associated recommendation method of claim 1, wherein associating a plurality of basic tags of the text information according to semantic information of the text information to obtain advanced semantic tags of the input information comprises:

judging whether the input information is input information input for the first time;

if the input information is not the input information input for the first time, acquiring a plurality of basic labels before the input information according to a preset acquisition rule;

associating a plurality of basic labels of the text information and a plurality of basic labels at the last moment according to the semantic information of the text information to obtain an advanced semantic label of the input information; the plurality of basic labels at the previous moment are a plurality of basic labels before the input information.

5. An associated recommendation device based on a tag knowledge graph is characterized by comprising:

the recommendation unit is used for acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the input information if the input information of the user is received so as to perform information pushing service on the user;

the first input unit includes:

the word segmentation unit is used for carrying out word segmentation processing on each text marked with a plurality of basic labels to obtain words of each text;

The second input unit is used for inputting the words of each text into a pre-trained language model to obtain semantic information of each text;

the first acquisition unit is used for associating a plurality of basic labels in each text according to the semantic information of each text so as to obtain a high-order semantic label of each text;

the recommendation unit includes:

a first judging unit configured to judge a type of the input information;

the conversion unit is used for converting the voice input information into text information of the input information and acquiring a high-order semantic tag matched with the input information from the tag knowledge graph according to the text information if the type of the input information is voice input information;

the conversion unit includes:

the second acquisition unit is used for acquiring semantic information of the text information and a plurality of basic labels of the text information;

and the third acquisition unit is used for associating a plurality of basic labels of the text information according to the semantic information of the text information so as to obtain the advanced semantic label of the input information.

6. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the tag knowledge-graph based associative recommendation method of any one of claims 1 to 4 when the computer program is executed.

7. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program, which when executed by a processor causes the processor to perform the tag knowledge-graph based associative recommendation method according to any one of claims 1 to 4.