KR20180066705A

KR20180066705A - Method and apparatus for analyzing vulnerability of learner

Info

Publication number: KR20180066705A
Application number: KR1020160167847A
Authority: KR
Inventors: 김채연
Original assignee: 김채연
Priority date: 2016-12-09
Filing date: 2016-12-09
Publication date: 2018-06-19

Abstract

The present invention provides a method for analyzing a vulnerability of a learner and providing a recommendation problem which provides a database of a problem solving result of a student and systematically classifies the result, thereby providing a reliable recommendation problem required for a learner. The method for analyzing a vulnerability of a learner and providing a recommendation problem of the present invention comprises: a DB generating step of providing a database of problem data, learner information data and problem solving data for each learner; a vector generating step of providing a vector of an incorrect answer accumulation value for each learner based on the generated DB; a cluster generating step of calculating a vector similarity between the learners, classifying the learners according to the calculated similarity and the number of learners, and generating a plurality of clusters; a cluster exploring step of exploring the cluster corresponding to an attribute of a subject to be analyzed; and a result providing step of providing a vulnerability and a recommendation problem to the subject to be analyzed based on the explored cluster.

Description

TECHNICAL FIELD The present invention relates to a method and apparatus for analyzing a vulnerability of a learner,

The present invention relates to a method and apparatus for analyzing a tendency of a learner to provide a recommendation problem according to an analysis result.

2. Description of the Related Art [0002] With the recent development of electronic devices and the Internet network, a system capable of providing educational contents on-line using an electronic book and an Internet network using electronic devices has been popular. By providing education contents online, there is no restriction on the paper, and it is possible to receive contents more easily.

In Korean Patent Laid-Open Publication No. 2002-0045431 published on Jun. 19, 2002, each problem of the question bank is divided into various criteria using a computer, and the level of difficulty and level of each standard is evaluated by the user, Technology. This system not only has the merit of allowing users to evaluate themselves, but it also enables classification by types of users by using many users of the question banks, and it is possible to manage the database of each user's problem type, Can be provided.

However, the existing system is only applied to the computer system in the form that has been continued in the real life so far, and it is hard to find any advantage as a teaching material and learning tool for level learning other than the advantage that the user is familiar. It is also possible to provide similar problems related to incorrect answers, but it does not provide similar reliability problems that are necessary in actual learner criteria. In addition, there is a problem that a system for classifying and providing similar problems in a systematic manner is not established, and thus a similar problem with high reliability can not be provided.

The present invention aims to provide a reliable recommendation problem that is necessary for a learner by classifying the results of a problem solving process of an actual student into a database and systematically classifying the results.

Further, the present invention can classify the clusters according to various criteria through the established database, and provide analysis results with improved fitness for the learner according to demand.

According to an embodiment of the present invention, a method for analyzing a learner's vulnerability and recommending a problem includes: DB generation of problem data, learner information data and problem solving data for each learner; vectorization of an incorrect answer accumulation value for each learner based on the generated DB A vector generation step, a vector generation step for calculating vector similarity between learners, a cluster generation step for generating a plurality of clusters by classifying learners according to the calculated degree of similarity and the number of learners, a cluster search step and a search step And providing a result to the analysis subject based on the clustering cluster.

According to an embodiment, the vector generation step includes selecting at least one category of problem attributes and setting at least one problem attribute.

According to the embodiment, the step of counting the vector dimension corresponding to the problem attribute in which the wrong answer occurs at the occurrence of the wrong answer by varying the vector dimension for each set problematic property.

According to an embodiment of the present invention, there is a step of accumulating the number of incorrect answers every time an error occurs and dividing the count value of each vector dimension by the accumulated number of incorrect answers.

According to the embodiment, the cluster generation step further includes a step of calculating a vector similarity by weighting each vector dimension.

According to the embodiment, the cluster searching step includes a step of extracting a center point representative of each cluster for each generated cluster.

According to the embodiment, the step of extracting the cluster having the largest degree of similarity is performed by calculating the degree of similarity between the extracted center point and the analysis subject.

According to the embodiment, the cluster generating step further includes classifying the learner by region or school to create a cluster.

According to another embodiment, the apparatus for providing a learner's vulnerability analysis and recommendation problem may include a vector generation unit for vectorizing the wrong answer accumulation values for each learner based on the DB and the DB for database of problem data, learner information data, A clustering section for classifying the learner into a plurality of clusters according to the calculated degree of similarity and the number of learners, a cluster search section for searching clusters having similar properties to the analysis target, And a result providing section that provides vulnerability and recommendation problems.

According to an embodiment of the present invention, the vector generation unit counts vector dimensions corresponding to a problem attribute in which an incorrect answer occurs when a wrong answer occurs, and counts the number of incorrect answers, Can be pressed.

According to the embodiment, the cluster search unit can calculate a center point representing each cluster for each generated cluster, and calculate the similarity between the calculated center point and the analysis target person, thereby extracting the cluster having the largest similarity.

The present invention includes a computer-readable recording medium storing a program for causing a computer to execute a method according to an embodiment of the present invention.

The present invention includes a program stored on a recording medium for causing a computer to execute a method according to an embodiment of the present invention.

According to the present invention, the results of problem solving by actual students are stored in a database, and the results are systematically classified, thereby providing a highly reliable recommendation problem required for a learner.

1 is a flowchart schematically illustrating a configuration of a method of analyzing a learner's vulnerability and providing a recommendation problem according to an embodiment of the present invention.
2 to 4 are diagrams illustrating a database according to an embodiment of the present invention.
5 is a diagram illustrating a cluster creation step according to an embodiment of the present invention.
6 is a diagram illustrating a center point of a cluster according to an embodiment of the present invention.
7 is a diagram illustrating a cluster search step according to an embodiment of the present invention.
FIG. 8 is a block diagram schematically showing a configuration of a learner vulnerability analysis and recommendation problem providing apparatus according to an embodiment of the present invention. Referring to FIG.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to clarify the technical idea of the present invention. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a block diagram of a computer system according to an embodiment of the present invention; Fig. For convenience of explanation, the apparatus and method are described together when necessary.

1 is a view schematically showing a configuration of a method of analyzing a learner's vulnerability and providing a recommendation problem according to an embodiment of the present invention.

Referring to FIG. 1, a method for analyzing a learner's vulnerability and providing a recommendation problem according to an exemplary embodiment of the present invention includes a DB generation step S1, a vector generation step S2, a cluster generation step S3, , And a result providing step S5.

In the DB creation step S1, problem data, learner information data, and problem solving data for each learner are converted into a database. As shown in FIG. 2, the problem data may include information about the problem content, difficulty, solving time, the expense ratio, and / or the problem solving ability for each problem. The learner information data may include the learner's region, school, grade, and / or gender information. Learner-specific problem solving data may mean information that each learner included in the learner information data solves the problem contained in the problem data, matching the learner, the problem, and the result thereof. For example, the problem data and the learner information data have hundreds, thousands or even tens of thousands of their data pools, so that sufficient learner-specific problem solving data can be databaseized.

In the vector generation step S2, the wrong answer accumulation value is vectorized for each learner based on the generated DB. For example, based on the learner-by-learner problem solving data of the DB, the case where the problem solving result is an incorrect answer for each learner is cumulatively calculated. At this time, by calculating an incorrect answer for each problem type, it can be expressed as a vector spatially representing an incorrect answer accumulation value. Since the vector is generated for each learner, the generated vector may correspond to an indicator indicating the propensity or attribute of each learner regarding the problem solving.

FIG. 3 and FIG. 4 are diagrams showing the configuration of a vector generation step (S2) according to an embodiment of the present invention. According to the embodiment, the vector generation step (S2) includes a step of selecting a category related to the problem attribute and setting at least one problem attribute. The user can select at least one of the categories constituting the problem data stored in the DB. When a category is selected, keywords constituting the content of the category are set as problem attributes. For example, as shown in FIG. 3, when two categories of the question data and the required ability are selected from the categories constituting the problem data, the keywords constituting the unit name (ex. Expression), and the keywords that make up the necessary competence (eg, understanding of terminology and internal linkage) can all be set as problem attributes.

Thereafter, the vector generation step S2 includes counting the vector dimension corresponding to the problem attribute in which the wrong answer occurs when the wrong answer is generated, by varying the vector dimension for each set problematic property. For example, if the problem attribute is set to 291 as shown in FIG. 4, the vector may have a vector dimension of 291 dimensions. In other words, the vectors may have corresponding dimensions for each problem attribute. However, the present invention is not limited to this, and the ratio of the problem attribute to the vector dimension may be N: M.

For example, if the attribute corresponding to the first dimension is a prime factor (a), the attribute corresponding to the second dimension is the greatest common denominator and the least common multiple (b), and the attribute corresponding to the third dimension is an integer and rational number (a, b, c, d, e, f). If a learner has a wrong answer to the problem of integers and rational numbers, the value of c in the learner's vector is counted.

Thereafter, the vector generation step S2 includes accumulating the number of incorrect answers every time an error occurs, and dividing the count value of each vector dimension by the accumulated number of incorrect answers. If the vector is generated by counting only the number of incorrect answers in different situations, it is not possible to determine the propensity between the learners with the corresponding vector. Accordingly, the number of incorrect answers generated by the learner is accumulated and counted for each learner regardless of the problem attribute every time an incorrect answer is generated. By dividing the value of each vector dimension by the total number of incorrect answers, and constructing each vector dimension value as an average value, the influence of the total problem solving number can be excluded when determining the tendency among the learners using the vector.

5 is a diagram illustrating a cluster generation step according to an embodiment of the present invention. In the cluster creation step S3, vector similarity is calculated between learners, and the learners are classified according to the calculated degree of similarity and the number of learners to generate a plurality of clusters. When the vector value is generated for each learner, the similarity can be calculated between the learner using the generated vector value. For example, the similarity may be a distance between learners (vector value distance), and may be a cosine similarity such as the following expression.

When the similarities are calculated between learners, one cluster is created by grouping the learners with similarity satisfying certain criteria. For example, if the similarity value corresponding to the criterion for creating one cluster is A, two learners belong to one cluster if the similarity degree between the two learners is A or more, and belong to different clusters if they are below A. Further, the similarity value corresponding to the criterion for creating the cluster may vary depending on the total number of learners or the number of clusters per cluster. For example, if the total number of learners is large, the number of people per cluster can be increased and the similarity criterion value can be lowered. Conversely, if the total number of learners is small, the number of people per cluster can be reduced and the similarity criterion value can be increased.

According to the embodiment, in the cluster creation step S4, a vector similarity degree can be calculated by assigning weights to vector dimensions. The weights can be set to assign weights to the selected categories when the category is selected from the problem data and the problem attributes are set. Alternatively, the attributes to be weighted can be individually weighted .

In the cluster search step (S4), the cluster corresponding to the attribute of the analysis subject is searched. The analysis subject may be one of the learners belonging to the learner information data provided in the DB, or may be a learner who is not related to the learner. The analysis subject is a learner who analyzes his / her vulnerability from clustered learner information from DB and wants to receive recommendation problem. The subject to be analyzed has a vector value that can spatially represent the attribute of the subject to be analyzed, like the learner provided in the DB. In the clustering search step S4, it is determined whether the clusters are similar to the clusters generated by using the vector values of the analysis target persons, thereby searching for clusters to which the analysis target persons can belong.

According to the embodiment, the cluster search step (S4) extracts a center point representative of each cluster for each cluster generated first. A center point can refer to a learner who is centered spatially in a cluster. Alternatively, the center point may mean a learner having an absolute average vector value among learners belonging to one cluster. If the center point is a representative value representing the attribute of each cluster, it can also be defined by a method other than the above-described method.

Then, in the cluster search step S4, the similarity between the extracted center point and the analysis target is calculated, and the cluster having the greatest similarity is extracted. The method of calculating the similarity between the extracted center point and the analysis target can be the same as the method of calculating the similarity between learners when creating a cluster.

In the result providing step S5, the vulnerability and the recommendation problem are provided to the analysis subject based on the detected cluster. If a cluster with a similar tendency to the subject of analysis is searched, the vulnerability and recommendation problem to be provided by the analysis subject based on the vulnerability and recommendation problem, It can be newly extracted. Thus, by providing the vulnerability and recommendation problem through the attribute of the current analysis subject and the attribute of the cluster having the similar tendency to the analysis target, analysis information about the problem solution can be more appropriately provided to the learner.

According to a further embodiment, the cluster creation step (S4) may further include classifying the learner by region or school and creating a cluster. In this case, rather than vectorizing the learner attribute according to the problem attribute, a cluster may be created based on the learner's region or school, and information about the problem attribute may be provided for each generated cluster.

The present invention can also be embodied as computer-readable codes on a computer-readable recording medium. The computer-readable recording medium includes all storage media such as a magnetic storage medium, an optical reading medium, and the like. It is also possible to record the data format of the message used in the present invention on a recording medium.

FIG. 8 illustrates a configuration of a learner vulnerability analysis and recommendation problem providing apparatus according to an embodiment of the present invention. 8, the learner's vulnerability analysis and recommendation problem providing apparatus includes a DB 10, a vector generating unit 20, a cluster generating unit 30, a cluster searching unit 40, and a result providing unit 50 do.

The DB 10 converts the problem data, the learner information data, and the problem solving data for each learner into a database. The vector generating unit 20 vectorizes the wrong answer accumulation value for each learner based on the DB 10. [ The cluster generating unit 30 calculates the degree of similarity between vectors among learners, and classifies learners into a plurality of clusters according to the calculated degree of similarity and the number of learners. The cluster search unit (40) searches clusters having properties similar to those of the analysis target. The result providing unit 50 provides the vulnerability and the recommendation problem to the analysis subject based on the detected cluster. The features of the respective components are the same as those of the learner's vulnerability analysis and recommendation problem providing method.

According to the embodiment, the vector generating unit 20 counts the vector dimension corresponding to the problem attribute in which an incorrect answer occurs when a wrong answer occurs, by varying the vector dimension for each problem attribute, accumulates the number of incorrect answers, Can be divided.

According to the embodiment, the cluster search unit 30 can calculate the center point representing each cluster for each generated cluster, and calculate the similarity between the calculated center point and the analysis target person, thereby extracting the cluster having the greatest similarity.

The present invention has been described in detail with reference to the preferred embodiments shown in the drawings. These embodiments are to be considered as illustrative rather than limiting, and should be considered in an illustrative rather than a restrictive sense. The true scope of protection of the present invention should be determined by the technical idea of the appended claims rather than the above description. Although specific terms are used herein, they are used for the purpose of describing the concept of the present invention only and are not used to limit the scope of the present invention described in the claims or the claims. Each step of the present invention need not necessarily be performed in the order described, but may be performed in parallel, selectively, or individually. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. It is to be understood that the equivalents include all components that are invented in order to perform the same function irrespective of the currently known equivalents as well as the equivalents to be developed in the future.

10: Database 20: Vector generating unit
30: Cluster Generation Unit 40: Cluster Search Unit
50: Results

Claims

A DB generating step of converting question data, learner information data, and learner-specific problem solving data into a database;
A vector generating step of vectorizing the wrong answer cumulative value for each learner based on the generated DB;
A cluster generating step of calculating a vector similarity degree among learners, and classifying learners according to the calculated degree of similarity and the number of learners, thereby generating a plurality of clusters;
A cluster search step of searching a cluster corresponding to the attribute of the analysis subject; And
And the result providing unit includes a result providing step of providing a vulnerability and a recommendation problem to the analysis subject based on the discovered cluster.

2. The method of claim 1,
Setting at least one problem attribute by selecting a category related to the problem attribute;
Counting a vector dimension corresponding to a problem attribute in which an incorrect answer occurs when a wrong answer is generated by varying a vector dimension for each set problem property; And
And accumulating the number of incorrect answers each time a wrong answer is generated and dividing the count value of each vector dimension by the accumulated number of incorrect answers.

3. The method according to claim 2,
And calculating vector similarity by assigning weights to vector dimensions. The method of claim 1,

2. The method of claim 1,
Extracting a center point representative of each cluster for each generated cluster; And
Calculating a degree of similarity between the extracted center point and the analysis target person, and extracting a cluster having the largest similarity degree.

5. The method according to claim 4,
And providing the vulnerability and the recommendation problem based on the incorrect information of the analysis target person and the discovered similar cluster.

2. The method of claim 1,
And classifying the learner by region or school to create a cluster. &Lt; Desc / Clms Page number 19 >

A computer-readable storage medium storing a computer program for implementing a method for analyzing a vulnerability of a learner and providing a recommendation problem according to any one of claims 1 to 6.

A DB for converting question data, learner information data, and learner-specific problem solving data into a database;
A vector generating unit for vectorizing an incorrect answer accumulation value for each learner based on the DB;
A cluster generating unit for calculating vector similarity between learners, and classifying learners into a plurality of clusters according to the calculated degree of similarity and the number of learners;
A cluster search unit for searching for clusters of attributes similar to the analysis subject; And
And a result providing unit for providing a vulnerability and a recommendation problem to the analysis subject based on the searched cluster.

9. The apparatus of claim 8, wherein the vector generation unit
Wherein a vector dimension corresponding to a problem attribute in which an incorrect answer occurs is counted by varying a vector dimension for each problem attribute, and a count value of each vector dimension is divided by the number of incorrect answers accumulated by accumulating the number of incorrect answers. And a recommendation problem providing apparatus.

9. The apparatus of claim 8, wherein the cluster search unit
Calculating a center point representative of each cluster for each generated cluster, calculating the similarity between the calculated center point and the analysis target, and extracting a cluster having the largest similarity,