KR102213482B1 - Method, apparatus and computer program for analyzing education contents and users - Google Patents

Abstract

The present invention relates to a data analysis service server that provides user-customized learning content, and includes at least one multiple-choice problem including at least one view for a specific subject, and includes the concept of the problem for each of the problems. A problem database containing information on degrees; And for each of the problems included in the problem database, estimating the degree of inclusion of the concept of the problem, providing the problem included in the problem database to a user device, and the user's response to the problem from the user device. Collecting view selection data, estimating the selection probability for each view for the view included in the problem, estimating the degree of understanding of the concept of the user using the user's view selection data, and identifying the concept that the user is vulnerable to. And a control unit for estimating and recommending a problem constructed including the estimated concept.

Description

A method, apparatus, and computer program for analyzing educational content and users of the above content {METHOD, APPARATUS AND COMPUTER PROGRAM FOR ANALYZING EDUCATION CONTENTS AND USERS}

The present invention relates to a method of analyzing the problem and the user. The present invention relates to a method of recommending a customized problem to each user based on the analysis data of the problem and the analysis data of the user.

Until now, educational content has been generally provided as a package. For example, a book of problems recorded on paper contains at least 700 questions per volume, and online or offline lectures are also sold in 1-2 hours, bundles the amount to study for at least a month.

However, from the perspective of educated students, there is a need for personalized content rather than a package form because both vulnerable units and vulnerable problem types are different. This is because it is much more efficient to select and study only the weak problem types of the weak unit than to solve the entire 700 problems in the problem book.

However, it is very difficult for students who are trainees to identify their own weaknesses. Furthermore, it is not easy to provide optimized problems to individual students because even in the conventional education industry such as academies and publishing houses, students and problems are analyzed based on subjective experience and intuition.

As described above, in a conventional educational environment, it is not easy for an educator to provide personalized content that can produce learning results most efficiently, and a problem arises that students quickly lose a sense of accomplishment and interest in packaged educational content.

Republic of Korea Patent Publication No. 10-2016-0117097 (2016.10.10)

The present invention aims to solve the above problems. More specifically, an object of the present invention is to provide a method and a computer program for analyzing problems and analyzing users. Further, an object of the present invention is to provide a method of extracting a customized problem for reinforcing each user's vulnerability from a problem database based on the analysis data of the problem and the analysis data of the user.

A data analysis service server providing user-customized learning content according to an embodiment of the present invention includes at least one multiple-choice problem including at least one view for a specific subject, and the concept of the problem for each of the problems. A problem database containing information on the degree of inclusion of; And for each of the problems included in the problem database, estimating the degree of inclusion of the concept of the problem, providing the problem included in the problem database to a user device, and the user's response to the problem from the user device. Collecting view selection data, estimating the selection probability for each view for the view included in the problem, estimating the degree of understanding of the concept of the user using the user's view selection data, and identifying the concept that the user is vulnerable to. And a control unit for estimating and recommending a problem constructed including the estimated concept.

According to an embodiment of the present invention, there is an effect of more precisely analyzing a problem and a user. Furthermore, according to an exemplary embodiment of the present invention, since a problem for each user can be extracted from a problem database based on the problem analysis data and the user analysis data, there is an effect of increasing user learning efficiency.

1 is a view for explaining the operation of the data analysis framework according to an embodiment of the present invention

The present invention is not limited to the description of the embodiments to be described below, and it is obvious that various modifications may be made without departing from the technical gist of the present invention. Further, in describing the embodiments, descriptions of technical contents that are widely known in the technical field to which the present invention pertains and are not directly related to the technical subject matter of the present invention will be omitted.

Meanwhile, in the accompanying drawings, the same elements are represented by the same reference numerals. In addition, some components in the accompanying drawings may be exaggerated, omitted, or schematically illustrated. This is to clarify the gist of the present invention by omitting unnecessary description not related to the gist of the present invention.

Recently, as the spread of IT devices is expanding, data collection for user analysis is becoming easier. If user data can be sufficiently collected, the user's analysis becomes more precise, and the most suitable form of content can be provided to the user.

Along with this trend, there is a high need for providing customized educational content, especially in the education industry. This is because students want to study problems with the highest learning effect and to study with high efficiency in a limited time.

For a simple example, if a user has poor understanding of <verb tense> in an English subject, learning efficiency will be higher if a problem including the concept of <verb tense> can be recommended. However, in order to provide user-customized educational contents in this way, precise analysis of each content and individual users is required.

The present invention is to solve the above problem, and a data analysis server according to an embodiment of the present invention may accurately analyze learning data by applying a machine learning framework to analysis of learning data.

For example, according to an embodiment of the present invention, a problem database is constructed, a problem set is constructed from the problem database, and the result data of the user's problem solving for the problem set is collected, based on whether the user is correct or incorrect. User modeling vectors can be created. The value of the user modeling vector may be interpreted as representing the characteristics of all users of the corresponding user.

Furthermore, by using the user modeling vector and/or problem modeling vector, the location of a specific user in all users, other users who can cluster into groups similar to a specific user, similarity between other users and the user, the location of a specific problem in the entire problem, Another problem that can be clustered into a group similar to the above problem, and the similarity between the problem and the other problem may be mathematically calculated. Furthermore, users and problems can be clustered based on at least one attribute.

In this case, it should be noted that in the present invention, the user modeling vector and the problem modeling vectors are limited and cannot be interpreted as to which attributes or features are included.

For example, according to an embodiment of the present invention, the user modeling vector may include a degree to which the user understands a certain concept, that is, a degree of understanding of the concept. Furthermore, the problem modeling vector may include what concepts the problem is composed of, that is, a conceptual diagram.

However, in order to analyze learning data, there are several problems to be solved.

The first is that if only the correct or incorrect answer as a result of problem solving is used as a parameter for generating the modeling vector, it is necessary to collect a large amount of problem solving result data in order to provide an analysis result with a certain degree of reliability or higher.

Multiple choice questions are composed of not only the text but also the viewing element. However, according to the conventional method of reflecting only correct or incorrect answers as a source of analysis, if two students make the same problem but choose a different option, the effect of the problem on the calculation of the user modeling vector of the two students will be the same.

In other words, if a student is wrong while selecting the view for <Verb Tense> and is wrong while selecting the view for <Verb Tense>, according to the conventional method, the vector value of the question is calculated. The effect of the solution result data is the same, and the solution result of the problem will not be sufficiently reflected in user analysis and will be substantially diluted.

Therefore, in order to obtain the result that the two students have poor understanding of the gerund and the tense of verbs, respectively, it will be necessary to collect a larger amount of problem solving result data to overcome the diluted result.

That is, according to the conventional method of using only the correct answer result as an input value for data analysis, a larger amount of problem solving result data must be collected in order to analyze a problem and a user with more than a certain range of reliability.

According to an embodiment of the present invention, the above problems can be solved. More specifically, according to an embodiment of the present invention, a user and/or a problem may be analyzed by applying a view parameter selected by the user from the problem solving result data. Through this, the same problem solving result data can be expanded as many as the number of views of the corresponding problem, and there is an effect that users and problems can be analyzed more precisely with the same result data.

1 is a flowchart illustrating a method of analyzing result data of problem solving according to an embodiment of the present invention.

Steps 110 and 1205 are steps of collecting learning data and modeling problems and/or users in a data analysis system according to an embodiment of the present invention.

According to an embodiment of the present invention, in step 110, solution result data for all problems and all users may be collected.

More specifically, the data analysis server may configure a problem database and collect result data of all users' solution for all problems belonging to the problem database.

Furthermore, the data analysis server may collect solution result data in a manner that collects the result of solving the problem by the user. The problem database includes listening evaluation problems and may be in the form of text, images, audio, and/or video.

At this time, the data analysis server may configure the collected problem solving result data in the form of a list of users, problems, and results. For example, Y (u, i) denotes the result of the user u solving the problem i, and may be assigned a value of 1 for the correct answer and 0 for the incorrect answer.

Furthermore, the data analysis server according to an embodiment of the present invention configures a multidimensional space composed of a user and a problem, and assigns a value to the multidimensional space based on whether the user has a problem or is wrong, so that each user and/or problem The modeling vector for can be calculated. (Step 120) At this time, the user modeling vector and/or the feature included in the problem modeling vector should be interpreted as not being specified.

The reason for modeling a user and a problem in the data analysis framework according to an embodiment of the present invention is ultimately to accurately predict whether a specific user will fit a specific problem or not.

However, as described above, the correct answer rate for the user's problem can be estimated by collecting and analyzing the user's correct answer results, but it can also be estimated using the selection probability of a choice.

For example, if the first user has a probability of selecting an option for a specific question (0.1, 0.2, 0, 0.7), the user will select option 4 with a high probability, and when the correct answer to the corresponding question is 4, the first Users can expect a high probability of hitting the problem.

Furthermore, the difficulty of the problem may be estimated by collecting and analyzing the correct and incorrect results of all users, but it may also be estimated using the selection probability of the example.

For example, for a specific problem composed of 4 choices, if the user's probability of selecting an option is (0.5, 0.1, 0.3, 0.6), the probability that users will choose option 1 and option 4 in the second question is Since they are similar, the second problem can be classified as having higher difficulty.

To this end, in the example of FIG. 1, the data analysis server may collect solution result data for all problems and all users, including view elements selected by the user. (Step 130) More specifically, the data analysis server may configure a problem database, and collect result data on which view all users have selected for all problems belonging to the problem database.

For example, the data analysis server builds a database for various problems on the market, and collects the solution result log by collecting the view selection result and which view selected by the user. You can also collect the correct and incorrect results for the question. In this case, the data analysis server may construct a list of users, problems, and views selected by the user in the problem.

Thereafter, the data analysis server can perform data analysis processing by expanding one problem into view units. (Step 140)

For example, if problem i is a multiple-choice problem, then problem i expands to 4 variables in (i, 1) (i, 2) (i, 3) (i, 4), and each question-view is selected by each user Whether or not can be given as a value.

For example, E (i, j) means view j of problem i, Y'(u, E) means whether user u has selected view j of problem i, according to an embodiment of the present invention. A value of 1 may be assigned in case of 1, and 0 in case of no selection.

Furthermore, the data analysis server configures a multi-dimensional space with the user and problem-view as variables, and assigns a value to the multi-dimensional space based on whether the user selects the problem-view, so that each user and/or problem-view The modeling vector for can be calculated. (Step 150)

The data analysis server according to an embodiment of the present invention may estimate a probability, that is, a selection rate, for an arbitrary user to select an arbitrary problem-view by using the user modeling vector and/or the problem-view modeling vector.

In this case, the selectivity may be estimated by applying various algorithms to the user modeling vector and/or the problem-view modeling vector, and the algorithm for calculating the selectivity is not limited in interpreting the present invention. For example, the view selection rate may be estimated by collecting selection data for a view of a corresponding problem with respect to other users, and calculating a ratio of selecting each view by other users.

As another example, according to an embodiment of the present invention, a user's problem-view selection rate may be estimated by applying a sigmoid function such as Equation 1 below. (x is the problem-view modeling vector, θ is the user modeling vector)

[Equation 1]

ｈθ(x) = 1 / (1 + e ^{(-θ * T * X)} )

As another example, the data analysis server according to an embodiment of the present invention may estimate a correct answer rate of a problem using the user's view selection rate.

For example, the correct answer rate of a question may be estimated using a statistical value of how many other users have answered the question.

On the other hand, for example, for a specific question composed of a four-point multiple-choice type, if a specific user's probability of selecting an answer is (0.5, 0.1, 0.3, 0.6) and the answer is number 1, what is the probability that the user will answer the question? Becomes a problem. That is, a method of estimating the correct answer rate for the corresponding question may be considered by using a plurality of selection rates for the corresponding question.

According to an embodiment of the present invention, as a simple method of reducing the choice of choices to the rate of correct answers to a question, a method of comparing the choices of the correct answers to the choices of all the choices may be considered. In this case, in the previous example, the percentage of correct answers for the user's question will be calculated as 0.5 / (0.5+0.1+0.3+0.6). However, when solving a problem, the user does not understand the problem by dividing it in units of views, but understands it in units of questions including the composition of the entire view and the intention to present the question, so the selection rate and the correct answer rate cannot be simply connected.

Accordingly, according to an embodiment of the present invention, the correct answer rate of a corresponding question can be estimated from the answer selection rate by averaging the selection rates of all the answers of a corresponding question and applying the averaged selection rates of the correct answers to the selection rates of all the answers.

In the previous example, if the selection probability of the views is (0.5, 0.1, 0.3, 0.6), when the average is averaged over all views, the scale may be changed to (0.33, 0.07, 0.20, 0.41). When the answer answer is number 1, the averaged selection rate of answer 1 may be estimated as 0.33, and the correct answer rate of the corresponding user to the question may be estimated as 33%.

Furthermore, the data analysis server according to an embodiment of the present invention may estimate a correct answer rate of a problem by using the user's problem-view selection probability, and through this, the user's understanding of a specific concept may be estimated.

Furthermore, the data analysis server can calculate the learning efficiency for a specific problem of a specific user. (Step 165)

For example, the data analysis server may calculate that a problem in which the user is likely to be wrong will have high learning efficiency for the user. As another example, the data analysis server may calculate that a problem in which a corresponding user has a problem modeling vector including a weak concept has high learning efficiency for the corresponding user. As another example, the data analysis server may calculate a learning efficiency for a specific problem of a corresponding user based on a selection rate for each view of the problem.

As described above, the data analysis server according to an embodiment of the present invention uses the modeling vector of the problem view unit and the user modeling vector to determine the probability of selecting a problem view unit and the correct answer probability of the corresponding user problem unit for each user. Can be calculated. Based on this, the data analysis server according to an embodiment of the present invention, when the user selects the corresponding view to the selection probability for each view of a certain problem of a specific user, when the average value of the correct answer rate change values of other questions is applied, the corresponding problem is Learning efficiency can be calculated.

For example, consider a case in which user A's specific problem a has a selection rate of a1 for the first view, a2 for the second view, a3 for the third view, and a4 for the fourth view. have.

At that time, assuming that user A has selected the first view for the problem, the modeling vector of user A and the modeling vector of all problems will be slightly changed by the application of the first view selection event of user A, and according to the changed modeling vector. The percentage of correct answers for each question will change somewhat. Accordingly, the data analysis server may calculate an average value of change in correct answer rate AVC_a1 of all questions by applying the virtual event for the user A's selection of the first view.

In the same way, the data analysis server may calculate AVC_a2 assuming that the user A has selected the second view, calculate AVC_a3 assuming that the third view is selected, and calculate AVC_a4 assuming that the fourth view is selected.

Thereafter, the data analysis server can calculate the learning efficiency E(A,a) through the user A's problem a through the following equation.

[Equation 2]

E(A,a)= a1* AVC_a1 + a2* AVC_a2 + a3* AVC_a3 + a4* AVC_a4

Furthermore, the data analysis server calculates E(A) for the entire problem, and can estimate that the learning efficiency is highest in the order of the problem with the highest E(A).

Furthermore, the data analysis server according to an embodiment of the present invention may recommend user-customized content. For example, the data analysis server may provide user-customized problems by extracting problems included in the problem database in the order of high learning efficiency to the corresponding user. (Step 170) That is, according to the embodiment of the present invention, a problem with the highest learning efficiency can be provided to an arbitrary user.

As another example, the data analysis server may recommend user-customized content using at least one of a user modeling vector, a problem-view modeling vector, a view selection rate, a correct answer rate, a user's conceptual understanding, and a problem conceptual diagram. .

For example, the data analysis service server can recommend in-depth courses on the concept.

As another example, the data analysis service server may recommend a problem configured to include the plurality of concepts to a user having a low correct answer rate for a problem including all of the plurality of concepts.

Furthermore, the service server may create a list of recommended problems for a corresponding user and provide them by sorting the remaining problems in the order of a lower correct answer rate for a specific user, excluding problems already solved by the user among the problem database set.

For example, if the value in row 1 of the user-problem correct answer rate P is [0.3, 0.4, 0.1, 0.9, 0.7], the probability that user 1 will answer questions 1, 2, 3, 4, and 5 is 30% and 40, respectively. It will be interpreted as %, 10%, 90%, 70%. The service server may provide the user with a problem recommendation list given priority in the order of problems 3, 1, 2, 5, and 4. At this time, if user 1 has already solved problem 5, the service server can create a problem recommendation list except for problem 5.

The embodiments of the present invention disclosed in the present specification and drawings are provided only to provide specific examples to easily explain the technical content of the present invention and to aid understanding of the present invention, and are not intended to limit the scope of the present invention. It is apparent to those of ordinary skill in the art that other modifications based on the technical idea of the present invention can be implemented in addition to the embodiments posted here.

Claims (4)

In the data analysis service server, in the method of providing customized learning content,
In the data analysis service server, a problem database including at least one multiple-choice problem including at least one example for a specific subject is configured, and the problem database includes, for each of the problems, the concept of the problem. Step a including information on the degree;
B, in the data analysis service server, providing the problem included in the problem database to a user device, and collecting view selection data of the user on the problem from the user device;
A step c of estimating, in the data analysis service server, a problem-view selection rate, which is a probability that the user selects a specific view of an arbitrary problem, using the view selection data of the user; And
In the data analysis service server, the user's question-view selection rate is used to estimate a correct answer rate for the user's arbitrary question, And d step of generating a list of recommended problems by giving priority to questions included in the problem database in the order of lower estimated correct answer rates, and providing the list of recommended problems to the user. . The method of claim 1, wherein step d is,
And averaging the question-view selection rate for the user's specific question over all the questions of the question, and estimating the averaged selection rate of the correct answer views of the question as the correct answer rate for the user. How to provide learning content. In the data analysis service server that provides user-customized learning content,
A problem database including at least one or more multiple choice questions including at least one example for a specific subject, and including information on inclusion of the concept of the problem for each of the problems; And
For each of the problems included in the problem database, the inclusion of the concept of the problem is estimated, the problem included in the problem database is provided to a user device, and the user's view of the problem from the user device Collecting selection data, estimating a problem-view selection rate, which is a probability that the user will select a specific view of an arbitrary problem, and estimating a correct answer rate for the user's random problem by using the user's problem-view selection rate, And a control unit for generating a list of recommended problems by giving priority to problems included in the problem database in an order of low estimated correct answer rate, and providing the list of recommended problems to the user. In the data analysis service server, in a computer program stored in a computer-readable recording medium to perform processing to provide user-customized learning content,
A function of constructing a problem database including at least one or more multiple-choice questions including at least one example for a specific subject, and estimating a degree of inclusion of the concept of the problem for each of the problems included in the problem database;
A function of providing the problem included in the problem database to a user device, collecting view selection data of the user for the problem from the user device, and estimating a selection rate for each view of the view included in the problem;
A function of estimating a problem-view selection rate, which is a probability that the user selects a specific view of an arbitrary problem, using the view selection data of the user; And
The user's question-view selection rate is used to estimate a correct answer rate for a random question of the user, and a list of recommended questions is generated by giving priority to questions included in the question database in the order of the estimated correct answer rate low, A computer program characterized in that it performs a function of providing the user with the recommended problem list.

Images