KR20210020718A - System for calculating study achievement and determining course class - Google Patents

Abstract

The system for calculating the academic achievement and determining the class according to an embodiment of the present invention relates to a system for calculating a student's academic achievement and determining a class in an academy operating a plurality of classes according to subject and level, and An input module that received the order of each subject; An order calculation module for calculating the order of the whole academy and within the class based on the order of each subject of the national practice test; A grade improvement achievement level calculation module for calculating the grade improvement achievement level of the student based on the data received from the input module and the grade level calculation module; And a class determination module for determining a class for each future subject of the student on the basis of the student's grade improvement achievement for each subject calculated by the grade improvement achievement calculation module.

Description

Academic achievement calculation and class determination system {SYSTEM FOR CALCULATING STUDY ACHIEVEMENT AND DETERMINING COURSE CLASS}

The present invention relates to a system for calculating academic achievement and class determination, and more specifically, to a system for calculating a student's academic achievement and determining a class in an academy that operates a plurality of classes according to subjects and levels.

Candidates take mock exams periodically to check their current skills, and based on the results of the practice exams, they identify subjects with insufficient skills and then establish a plan for future study methods.

In the case of the practice test conducted nationwide, it is possible to check the national ranking and percentage, and based on this, the ranking and percentage can be calculated even within the academy to which they belong.

A general academy operates a plurality of classes for each subject, and these classes are classified based on students' ability. For example, a class for any one subject can be divided into A, B, and C classes. , Here, Class A may be designated as Advanced Class, Class B as Normal, and Class C as Basic.

In the case of class A, which is an in-depth counterattack, more time is spent because more advanced classes are conducted and more advanced problems are solved compared to other classes. Therefore, there is not much time spent on solving problems.

In addition, students can change classes in consideration of their skills, and specifically, after the practice test, the class is changed based on the grades of the practice test taker.

On the other hand, the achievement level of individual subjects differs from student to student.For example, if student A's mock test grade is determined to be grade 3 in mathematics, grade 1 in Korean, and grade 4 in English, what are the subjects that A student lacks and has sufficient skills? You will be able to grasp cognition.

Even subjects with sufficient proficiency require sufficient study in order not to make mistakes in the actual SAT exam, and subjects with insufficient proficiency require sufficient study to raise them to sufficient proficiency.

On the other hand, when allocating study time, students tend to allocate study time emotionally rather than reasonably. This is because the student who is a party is more inclined to establish a plan through subjectivity rather than objectively grasping the current situation. There is a limit to planning.

Students naturally tend to devote a lot of study time to the subject they are good at, and this is because if a lot of study time is put into a difficult subject, the progress will be slow and this will hinder the progress of other subjects.

Therefore, a system and method are required to reasonably and effectively establish a study plan such as class and study time for each subject.

On the other hand, the following prior art document only discloses the contents of a school management and textbook ordering system using a wireless terminal, but does not disclose the technical gist of the present invention.

Korean Patent Application Publication No. 2003-0075065

A system for calculating academic achievement and determining a class according to an embodiment of the present invention aims at the following solutions in order to solve the above-described problems.

It provides a system that can objectively and reasonably establish a study plan such as class decision and study time allocation based on the grasp of the student's academic achievement in consideration of the student's current situation.

The problem of the present invention is not limited to those mentioned above, and other problems that are not mentioned will be clearly understood by those of ordinary skill in the art from the following description.

The system for calculating the academic achievement and determining the class according to an embodiment of the present invention relates to a system for calculating a student's academic achievement and determining a class in an academy operating a plurality of classes according to subject and level, and An input module that received the order of each subject; An order calculation module for calculating the order of the whole academy and within the class based on the order of each subject of the national practice test; A grade improvement achievement level calculation module for calculating the grade improvement achievement level of the student based on the data received from the input module and the grade level calculation module; And a class determination module for determining a class for each future subject of the student on the basis of the student's grade improvement achievement for each subject calculated by the grade improvement achievement calculation module.

The grade improvement achievement calculation module includes the data received from the input module and the ranking calculation module, the difficulty of the current class to which the student belongs, the study input time of the subject in the current class to which the student belongs, and the current class to which the student belongs. Calculate the student's grade improvement achievement based on the first characteristic data including the rank of each subject in the national practice test of other students, the rank of the whole academy of other students in the current class to which the student belongs, and the rank in the class. It is desirable to do.

It is preferable that the achievement level calculation module and the class determination module determine the student's achievement level and class for each subject based on a preset algorithm.

The algorithm is implemented as a deep learning algorithm composed of a neural network, the neural network network comprising: an input layer receiving the first feature data; At least one hidden layer configured to receive the first feature data from the input layer and include a plurality of nodes; And an output layer for determining the student's achievement level and class for each subject based on the operation result of the hidden layer, wherein the node defines a function that is a correlation between an input factor transmitted to the node and an output factor output from the node. It is preferable that the function includes a weight and a bias.

A storage module in which big data including changes in the academic achievement level of students of the academy, changes in the class, the ranking of each subject in the national practice test, the ranking in the academy, and the ranking in the class are stored; further includes, and the weight And a bias is preferably determined through machine learning performed on a neural network using big data stored in the storage module as training data.

Further comprising a training data generation module for generating training data for determining the weight and bias, wherein the training data generation module receives the first characteristic data and the student of the student based on a pre-learned criterion. A discrimination unit to determine the achievement level and class for each subject; And a generator that generates second feature data corresponding to the first feature data, wherein the discrimination device generates the discrimination criteria through learning using a Generative Adversarial Network (GAN). desirable.

The discriminating device is learned so as to distinguish the first feature data and the second feature data, and the generating device is learned so that the discriminating device cannot discriminate the first feature data and the second feature data, and the second feature data It is desirable to produce.

The system for calculating academic achievement and class determination according to an embodiment of the present invention utilizes input factors such as the student's national practice test scores, the rank in the academy, the rank in the class, the difficulty of the class, and the time spent studying in the class. There is an effect that it is possible to establish a more accurate and efficient student study plan by grasping the achievement level of grade improvement and further determining the future class.

The effects of the present invention are not limited to those mentioned above, and other effects not mentioned may be clearly understood by those of ordinary skill in the art from the following description.

1 is a block diagram schematically illustrating a system for calculating academic achievement and determining a class according to an embodiment of the present invention.
2 is a diagram schematically illustrating a neural network network used in a system for calculating academic achievement and determining a class according to an embodiment of the present invention.
3 and 4 are diagrams for explaining a training data generation module in a system for calculating academic achievement and determining a class according to an embodiment of the present invention.

A preferred embodiment according to the present invention will be described in detail with reference to the accompanying drawings, but the same or similar elements are assigned the same reference numerals regardless of the reference numerals, and redundant descriptions thereof will be omitted.

In addition, in describing the present invention, when it is determined that a detailed description of a related known technology may obscure the subject matter of the present invention, a detailed description thereof will be omitted. In addition, it should be noted that the accompanying drawings are only for easily understanding the spirit of the present invention and should not be construed as limiting the spirit of the present invention by the accompanying drawings.

Hereinafter, a system for calculating academic achievement and determining a class according to an embodiment of the present invention will be described with reference to FIGS. 1 to 4.

The system for calculating the academic achievement and determining the class according to an embodiment of the present invention relates to a system for calculating the academic achievement and determining the class of a student in an academy that operates a plurality of classes according to subjects and levels, as shown in FIG. It is configured to include an input module 100, a ranking calculation module 200, a grade improvement achievement calculation module 300, a class determination module 400, and a storage module 500.

The input module 100 is configured to receive the student's ranking and percentage for each subject of the national practice test, and the ranking calculation module 200 is based on the ranking and percentage of each subject of the student’s national practice test received from the input module 100. This is a configuration that performs the function of calculating the rank and percentage of the entire institute and within the class.

The grade improvement achievement calculation module 300 performs a function of calculating a student's grade improvement achievement level based on the data received from the input module 100 and the rank order calculation module 200 described above, and the class determination module 400 It performs a function of determining a student's future class for each subject based on the student's achievement for each subject calculated by the grade improvement achievement calculation module 300.

In particular, the grade improvement achievement calculation module 300 may calculate the grade improvement achievement of a student based on the first characteristic data, where the first characteristic data is transmitted from the input module 100 and the ranking calculation module 200 The data received, the difficulty of the current class to which the student belongs, the time spent studying the subject in the current class to which the student belongs, the ranking of each subject in the national practice test of other students in the current class to which the student belongs, and other students in the current class to which the student belongs Including the rankings within the entire school/class.

That is, by using the above-described first characteristic data as an input parameter, a student's achievement level for improvement of grades may be calculated through an algorithm to which a weight for each corresponding input parameter is applied, and a class for each subject of the student may be determined based on this.

In particular, in the system for calculating academic achievement and determining classes according to an embodiment of the present invention, the above-described algorithm is implemented as a deep learning algorithm composed of a neural network 10 as shown in FIG. Effective student achievement can be calculated and classes can be determined.

This neural network network 10 is configured to include an input layer 11, a hidden layer 12, and an output layer 13, as shown in FIG.

The input layer 11 performs a function of receiving the above-described first feature data, that is, an input variable, and the hidden layer 12 receives the first feature data from the input layer 11 and performs an operation.

It is preferable that a plurality of hidden layers 12 are provided, and each hidden layer 12 includes a plurality of nodes as shown in FIG. 2.

In each node, a linear or nonlinear function, which is a correlation between an input factor delivered to a corresponding node and an output factor output from the node, is defined, and the function is configured to include a weight and a bias.

For example, when the input factor at the node is defined as x and the output factor at the node is defined as y, the function defined at the node can be defined as "y=w*x+b". , Here, w means weight, and b means bias.

In the case of such a deep learning algorithm, optimal weights and biases are generated through machine learning, and by using them, accurate academic achievement and class determination are possible, which will be described later in detail.

The output layer 13 performs a function of determining a student's achievement level and class for each subject based on the operation result of the hidden layer.

On the other hand, the storage module 500 of the academic achievement calculation and class determination system according to an embodiment of the present invention is a change in achievement level of academic achievement of students in academy, a change in class, a ranking for each subject in the national practice test, and a ranking in the academy , This is a configuration that stores big data including the ranking in the class.

That is, the weights and biases of each node included in the neural network 10 described above are obtained through machine learning performed on the neural network 10 by using the big data stored in the storage module 500 as training data. Is determined.

On the other hand, what is needed in machine learning is to perform learning using a large amount of data above all, and to secure data, it is necessary to select data suitable for learning from among numerous data and connect it with the result of learning.

Deep learning is a technology based on such learning, and as a supervised learning method based on a number of data, there are convolutional neural networks (CNNs) and reccurrent neural networks (RNNs).

However, in the case of the above-described supervised learning method, since it is not easy to apply because it requires information data for a number of correct answers, if the training data is not sufficiently provided in the storage module 500 for learning the neural network 10 There is a problem in that it may be difficult to calculate elaborate weights and biases because machine learning is not properly performed.

Therefore, interest in unsupervised learning technology that is performed without labeling data is increasing, and in order to overcome the above-described problems, the system for calculating academic achievement and class determination according to an embodiment of the present invention does not provide information on the correct answer. Academic achievement can be calculated and classes can be determined through learning using Generative Adversarial Network (GAN), a technology that generates real data.

In the case of such a generative confrontation network, in order to create data and make it indistinguishable from real data, a role (generator) must be given to make the data a counterfeit and a role (discriminator) to distinguish whether this data is counterfeit or not.

Looking at this generative confrontation network in more detail, it is composed of a generator (1), a database (DB) (2), and a discriminator (3), as shown in FIG. As a result, generator 1 and discriminator 2 are networks composed of neural networks.

The discriminator (3) receives data from the database (2) or generator (1), and learns to determine the data (data_real) provided by the database (2) and the data (data_fake) provided from the generator (10) among these data. do.

On the other hand, the generator 1 learns to determine the data (data_fake) provided to the discriminator 3 as data_real, that is, the generator 1 learns so that the discriminator 3 determines the data_fake as data_real.

To summarize the learning directions of the two neural network learners (discriminator and generator), the discriminator (3) learns to distinguish between the actual data and the data generated by the generator (1), and the generator (1) is the discriminator (3). Learn to generate data that cannot be distinguished.

The generator 1 can generate similar data to the actual data stored in the database 2 by the repeated learning as described above, and thus the discriminator uses a large amount of training data generated by the generator 1 Since it can be secured, it is possible to produce sophisticated results.

The system for calculating academic achievement and determining class according to an embodiment of the present invention may further include a training data generation module that generates training data for determining the weights and biases described above, and such training data generation module is a generative confrontation. A network is applied, which will be described in more detail with reference to FIG. 4.

As shown in FIG. 4, the training data generation module inputs random noise data (Input_Z) to the generation unit 22. Initially, noise data (Input_Z) and a storage module are used for faster and faster reinforcement learning. The conventional first feature data (Real_Data) stored in 400 can be combined and input to the generator 22, through which the generator 22 can generate second feature data, which is a plurality of diversified feature data. I can.

The determination unit 21 may be the neural network 10 described above, receives the inputted first feature data (Real_Data) and second feature data (Fake_Data), determines the authenticity of them, and determines the result of Fake/Real. And the result is provided to the generation unit 22 so that the generation unit 22 can learn.

That is, the generation unit 22 improves the neural network so that the determination unit 21 determines the real data based on the authenticity determination of the determination unit 21, and eventually, as the learning of the generation unit 22 continues, the generation unit 22 ) Is to improve the neural network so that the determination unit 21 determines that it is real data.

In summary, the determination unit 21 learns to distinguish the first feature data and the second feature data, and the generation unit 22 allows the determination unit 21 to distinguish between the first feature data and the second feature data. It is learned so that the second feature data is generated.

Through the above process, the determination unit 21 can more accurately calculate academic achievement and determine the class, and through reinforcement learning using a generative confrontation network between the determination unit 21 and the generation unit 22 After sufficient learning of the determination unit 21 is completed, the generation unit 22 may be removed.

The embodiments described in the present specification and the accompanying drawings are merely illustrative of some of the technical ideas included in the present invention. Accordingly, it is obvious that the embodiments disclosed in the present specification are not intended to limit the technical idea of the present disclosure, but to explain the technical idea, and thus the scope of the technical idea of the present disclosure is not limited by these embodiments. Modified examples and specific embodiments that can be easily inferred by those of ordinary skill in the art within the scope of the technical idea included in the specification and drawings of the present invention are included in the scope of the present invention. It will have to be interpreted.

100: input module
200: rank order calculation module
300: Grade improvement achievement calculation module
400: class determination module
500: storage module

Claims (5)

In a system for calculating student achievement and class determination in academy that operates a plurality of classes according to subject and level,
An input module receiving the student's ranking for each subject of the national practice test;
An order calculation module for calculating an order in the whole academy and within the class based on the order in each subject of the national practice test;
A grade improvement achievement level calculation module for calculating the grade improvement achievement level of the student based on the data received from the input module and the grade calculation module; And
A class determination module for determining a class for each future subject of the student based on the student's grade improvement achievement for each subject calculated by the grade improvement achievement calculation module;
Academic achievement calculation and class determination system, including.
The method according to claim 1,
The grade improvement achievement calculation module includes the data received from the input module and the ranking calculation module, the difficulty of the current class to which the student belongs, the study input time of the subject in the current class to which the student belongs, and the current class to which the student belongs. Calculate the student's grade improvement achievement based on the first characteristic data including the rank of each subject in the national practice test of other students, the rank of the whole academy of other students in the current class to which the student belongs, and the rank in the class. Academic achievement calculation and class determination system.
The method according to claim 2,
The grade improvement achievement calculation module and the class determination module are a system for calculating academic achievement and class determining the achievement level of the student and the class for each subject based on a preset algorithm.
The method of claim 3,
The algorithm is implemented as a deep learning algorithm composed of a neural network network, wherein the neural network network,
An input layer receiving the first feature data;
At least one hidden layer configured to receive the first feature data from the input layer and include a plurality of nodes; And
An output layer that determines the student's achievement level and class for each subject based on the operation result of the hidden layer;
Including,
In the node, a function, which is a correlation between an input factor transmitted to the node and an output factor output from the node, is defined, and the function includes a weight and a bias.
The method of claim 4,
A storage module storing big data including changes in the academic achievement level of the students of the academy, changes in the class, the ranking of each subject in the national practice test, the ranking in the academy, and the ranking in the class; further includes,
The weight and bias are determined through machine learning performed on a neural network network using big data stored in the storage module as training data.

Images