JP2006072122A - Learner acquisition characteristic analysis system, its method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a learner acquisition characteristic analysis system to ask each learner for a score, deviation value, and affiliate type name, result information, such as teaching materials automatically selected in compliance with the affiliate type, his or her own acquisition tendency of each learner, and information on the acquisition tendency of the affiliate type near each learner and the characteristic of item error, and its method and program. <P>SOLUTION: The acquisition characteristic analysis system comprises a amount of the characteristics calculation section which outputs the performance of each item unit in the each learner as the amount of the characteristics, an average characteristics calculation section which calculates the average value of the performance for every item, and outputs the average value as the average amount of the characteristics, and a cluster extraction section which performs clustering of the each learner in a prescribed learner population based on the the average amount of the characteristics and the amount of the characteristics. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a learner acquisition characteristic analysis system, a method, and a program for analyzing acquisition characteristics for each learner in order to set appropriate teaching materials and learning methods for each learner.

Conventionally, in school education and e-learning, as a means of educational evaluation, tests that are performed regularly and irregularly in classes are used as units or semester breaks.
Here, the result data of the test is obtained as answer data such as an option number of each item (question) of each learner, scores, and correct / incorrect information.
In this test, the teacher asks questions related to the class, obtains response information for each learner, confirms the learner's acquisition status, and determines the grade.
That is, the score of the test is useful as summary information that expresses the acquired amount, and is used as reference information for determining a student's grade.

In addition, the teacher confirms the effect of the lesson contents until the test is performed based on the response information, reflects and improves the teaching method and lesson contents, and adjusts the lesson contents thereafter.
On the other hand, by seeing this response information, the learner confirms his / her achievement level with respect to the test contents, his / her test result in the entire learner, and his / her own performance for each item, Based on this information, we will consider future learning measures.

The following information is obtained as the test result.
a. As the degree of achievement of the test contents, the correct answer rate and score of the entire test and the average score of the entire learner.
b. As the degree of achievement for each item, the score for each item and the average score for each item of the entire learner.
c. As an individual's position within the entire learner,
・ Statistical information and score distribution such as average score and variance for all learners, rank and deviation value of each learner in the statistical information and score distribution ・ Average score of all learners in each item for achievement level for each item ( The score information of each learner is compared with “10 points on a scale and 6 points on average”). In addition, the distribution of responses (such as the selection rate of options) for each item of all learners is compared with the score information of each individual learner.

Using each piece of information described above, the teacher basically gives guidance to each learner on items with a low score (correct answer rate).
Here, based on the past learning history, learners are classified into types, and teaching materials that are considered useful for each learner are presented in correspondence with the results of each learner's test (see Patent Document 1). reference).
JP 2003-131549 A

However, in the learning course design support device shown in Patent Document 1, it is necessary to present different effective teaching materials for each learner to all learners, and the learning level of each learner A huge variety of materials must be prepared in advance.
When the number of teaching materials that can be prepared is small, there is a high possibility that it will not be used as a learning measure, but it is difficult to prepare appropriate teaching materials in advance for all learners.
For this reason, in order to properly prepare the corresponding teaching materials, there is “what kind of instruction is needed for learners” and information that can clearly grasp “what kind of teaching materials” should be prepared. Must be provided to teachers.

In other words, since it is a test to confirm the level of acquisition of the content in the class, if the affiliation type of a learner group (group number is unknown) that shows a learning tendency biased (out of average) from the test result can be extracted, It is clear what kind of instruction is needed for the learner.
Furthermore, if there is a means for analyzing the state of bias of each learner and the cause of the bias with respect to the average learning level, it is possible to obtain a guideline indicating what kind of teaching material is necessary.

In addition, for each learner, grasp the tendency of achievement in learning and how to make mistakes for each item, prioritize items to be strengthened, and take measures to eliminate the factors. However, it is necessary to eliminate the above-mentioned bias in learning.
The learning effect is higher when learning is actively promoted by recognizing the situation where the person is placed than when the teacher points out and passively learns about the indication.

In order to make it easy for each learner to confirm the achievement tendency of his / her level of achievement, feedback of the test results, such as scores, deviation values, affiliation type names, and teaching materials automatically selected according to the affiliation type Get as.
Furthermore, in order for each learner to recognize his / her situation, not only the result information described above, but also each learner compares his / her own learning tendency with that of other learners, It is necessary to feed back information from multiple points of view, such as the learning tendency of the affiliation type close to, and the characteristics of item mistakes.

  The present invention was made in view of such circumstances, for each learner, score information, deviation value and affiliation type name, result information such as teaching materials automatically selected according to this affiliation type, and each It is an object of the present invention to provide a learner acquisition characteristic analysis system, a method and a program for a learner to obtain his / her own acquisition tendency and information on the acquisition tendency of an affiliated type close to each learner and the characteristics of item errors.

  The learner acquisition characteristic analysis system of the present invention includes a characteristic amount calculation unit that outputs the achievement level of each item unit for each learner as a characteristic amount, calculates an average value of the achievement level for each item, and calculates the average characteristic amount An average characteristic amount calculation unit to output, and a cluster extraction unit that clusters each learner in a predetermined learner group based on the average characteristic amount and the characteristic amount.

  In the learner acquisition characteristic analysis system according to the present invention, the cluster extraction unit compares the average characteristic amount with the characteristic amount of all learners at the start of clustering, and has a characteristic amount equal to or higher than a predetermined similarity. Is output as an average cluster.

  The learner acquisition characteristic analysis system of the present invention includes a database storing attribute information including past learning histories of each learner, grades of subjects taken, good and bad subjects, and subjects of interest, and each extracted cluster And a cluster attribute extracting unit that extracts attribute information common to learners belonging to the cluster from the database.

  The learner acquisition characteristic analysis system of the present invention performs a test of a target learner to be estimated when a plurality of tests are performed in time series in a predetermined period and clustering is performed based on the test results of the plurality of stages. By comparing the characteristic quantity and the characteristic quantity of the comparison learner's test at the same stage in the cluster created in the past, and extracting the cluster containing the comparative learner with high similarity with the target learner, It has an affiliated cluster estimation unit for estimating a cluster included in the future when the target learner learns in the current state.

  In the learner acquisition characteristic analysis system of the present invention, for each cluster extracted by the similarity, the belonging cluster estimation unit includes attribute information common to each comparison learner belonging to the cluster, and attribute information possessed by the estimation learner. A comparison is made, and a cluster having a high degree of coincidence of attribute information is output.

  The learner acquisition characteristic analysis method of the present invention is a characteristic quantity calculation process in which the achievement level of each item in each learner is output as a characteristic quantity, and the average value of the achievement degree for each item is calculated, and the average characteristic quantity And a cluster extraction process in which each learner is clustered in a predetermined learner group based on the average characteristic amount and the characteristic amount.

  In the learner acquisition characteristic analysis method of the present invention, in the cluster extraction process, the average characteristic amount is compared with the characteristic amounts of all learners at the start of clustering, and the learner has a characteristic amount equal to or higher than a predetermined similarity. Is output as an average cluster.

  The learner acquisition characteristic analysis method of the present invention is extracted from a storage process of storing attribute information including past learning history of each learner, grades of subjects taken, good and bad subjects, and subjects of interest in a database. Further, each cluster has a cluster attribute extraction process in which attribute information common to learners belonging to the cluster is extracted from the database.

  The learner acquisition characteristic analysis method of the present invention performs a test of a target learner to be estimated when a plurality of tests are performed in time series in a predetermined period and clustering is performed based on the test results of the plurality of stages. By comparing the characteristic quantity and the characteristic quantity of the comparison learner's test at the same stage in the cluster created in the past, and extracting the cluster containing the comparative learner with high similarity with the target learner, It has a belonging cluster estimation process in which a cluster included in the future is estimated when the target learner learns in the current state.

  The program of the present invention is a characteristic amount calculation process in which the achievement level of each item in each learner is output as a characteristic amount, and the average value of the achievement level for each item is calculated and output as an average characteristic amount A computer that performs a learner acquisition characteristic analysis process having a characteristic amount calculation process and a cluster extraction process in which each learner is clustered in a predetermined learner group based on the average characteristic amount and the characteristic amount is executable. Program.

  In the cluster extraction process, the average characteristic amount is compared with the characteristic amounts of all learners at the clustering start time in the cluster extraction process, and learners having characteristic amounts equal to or higher than a predetermined similarity are output as average clusters. A program that can be executed by a computer that performs a learner acquisition characteristic analysis process.

As described above, according to the present invention, the cluster extraction unit acquires a plurality of groups as achievement levels of each item from a predetermined learner group, including clusters having average characteristics in the learner group. Because it is extracted according to the characteristics, the average group that each item has achieved the acquisition target on average in the class, and other learning characteristics have deviations from this average group and deviations that have an unknown tendency Each group can be extracted.
As a result, according to the present invention, by comparing the average group and the deviation group (single or plural), the characteristics of each group are analyzed, and the teaching and teaching method corresponding to the tendency of each group The effect that it is possible to obtain the guideline is obtained.

Further, according to the present invention, the learning characteristics of an average group can be obtained in a state in which a deviation group having a deviation from this group is separated and learners having greatly different learning characteristics are excluded. By analyzing the acquisition characteristics of the target group, it is possible to determine whether or not the course progress has a tendency intended by the teacher.
Furthermore, according to the present invention, it is possible to determine how suitable the lesson progress is for this learner group by obtaining the ratio of learners belonging to the average group in the learner group.

Further, according to the present invention, the cluster attribute extraction unit extracts attribute information common to learners included in each cluster for each extracted cluster, and outputs this as cluster attribute features.
As a result, according to the present invention, the prior knowledge common to the learners belonging to each cluster (the subjects taken in the learning history, the grades of each subject) and the learning environment (the subjects that are not good at the subject, subjects that are interested, etc.) are known. It is possible.
In addition, according to the present invention, the affiliation cluster estimation unit is performed during a predetermined period (semester or year), that is, at intervals within the predetermined period (for example, at the end of one semester of a one-year class). In the middle of the stage, when moving in the current learning situation, it is possible to extract which cluster (cluster in which past learners are clustered) can be finally extracted, so if the extracted cluster is other than the average group For example, the teacher can correct the bias of the learner's acquisition characteristics at an earlier time, and can guide the teacher to approach the average group.

Hereinafter, a learner acquisition characteristic analysis system according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of the embodiment.
In this figure, the characteristic quantity calculation unit 1 performs a test (by subject) periodically performed in time series during one semester for each learner belonging to a predetermined learner group, for example, a predetermined class of a university. The achievement level of each item is calculated from the inputted test response information, and is output as a characteristic amount of the learning characteristic of each learner.

  Here, the degree of achievement is based on the score of each item (for each question of each test) or the correct answer rate, for example, the answer information represented by a correct answer rate of 5 points out of 10 or 70%, The correct answer rate is standardized for each item. At this time, 5 out of 10 points is 0.5, and the correct answer rate of 70% is 0.7. The test answer information includes the results (scores and answer information) for each learner, and the contents and metadata of each item of the test performed.

Further, the characteristic amount calculation unit 1 writes the learner attribute information of each learner input together with the test answer information in the learner information database 5 in association with the identification number of each learner together with the above characteristic amount. At this time, the characteristic amount calculation unit 1 may obtain a deviation value of each learner and write it in the learner information database 5 together with the attribute information.
Here, the learner attribute information includes the history of subjects that have been taken (learned), the score or grade of the subjects that have been taken, information about the strengths and weaknesses of the subjects (information in which the level of strength or weakness is quantified), interest Some subject information (information in which the degree of interest is quantified) is included.

The average characteristic amount calculation unit 2 obtains the average value of all learners in the learning group for each item of the input test response information, and uses the average value of each item as the average characteristic amount as described above. It is calculated by the same method.
The cluster extraction unit 3 obtains the respective similarity by comparing the average characteristic amount and the characteristic amount of each learner, and the comparison of the characteristic amount between the learners, and learning having a similarity greater than or equal to a predetermined threshold value. Clustering is performed by extracting a person, and a cluster having the same tendency of acquisition characteristics is extracted.

The upper, middle, and lower cluster identification unit 15 obtains the score and achievement rate of each item in the test of input test response information as the achievement level, and accumulates the achievement level for each learner in all tests. By calculating the total achievement level, the top 30% of the entire learner group test is classified as the top cluster, the bottom 30% as the bottom cluster, and the other middle 40% as the grade group of the middle cluster. Cluster each learner by grade.
The upper, middle, and lower cluster identification unit 15 stores the obtained upper cluster, middle cluster, and lower cluster cluster acquisition characteristics and the learner identification numbers included in the clusters for each cluster.

The cluster attribute extraction unit 4 averages the characteristic amount of the learner's acquisition characteristic (feature vector of the acquisition characteristic) included in each cluster for each item, and uses this average result as the cluster acquisition characteristic, corresponding to each cluster. In addition, it is stored in the cluster characteristic database 6.
Further, the cluster attribute extraction unit 4 reads out learner attribute information of learners included in each cluster from the identification information of each learner, and attribute information common among learners in each cluster, for example, a subject subject is mathematical. And the results are extracted, the extracted attribute information is compared between the clusters, and the attribute information having a difference exceeding a predetermined attribute threshold from the attribute information extracted in other clusters is used as a cluster attribute feature. The data is stored in the cluster characteristic database 6 in association with each cluster.

  In addition, the item characteristic calculation unit 7 determines the content of each item (problem content, for example, a text sentence of the problem) and the metadata of this item (the learning target in the guidance point to be confirmed in this item) from the input test answer information. Information) and item characteristics (problem difficulty and discrimination level, etc.) are extracted over all items of all tests and stored in the item information database 8 for each item.

  The affiliation cluster estimation unit 9 calculates a feature amount in a predetermined period (for example, one semester period), so that a plurality of tests are performed in time series at each end of each learning target. Therefore, when each learner is clustered by the feature amount within the above period, the characteristic amount of the test of the target learner to be estimated and the comparison learner test at the same stage in the cluster created in the past Are compared with each other, and a cluster including a comparative learner having a high degree of similarity with the target learner is extracted, so that a cluster included in the future is estimated when the target learner learns in the current state.

  In addition to the above-described configuration, the learner learning characteristic analysis system according to the present embodiment includes a learning characteristic display unit 10 that displays a cluster learning characteristic of each cluster and a characteristic amount of a specific learner, and a cluster attribute characteristic of each cluster. The cluster attribute feature display unit 11 for displaying the item response status display unit 12 for displaying the response distribution for each cluster of the specified item, the item characteristics of each item, and the achievement level of each item of the specified learner at the same time It has an item achievement status display unit 13 for displaying, and an item information display unit 14 for displaying the contents and metadata of designated items.

<Clustering processing>
Next, the operation of the cluster extraction unit 3 of this embodiment will be described with reference to FIGS. Here, FIG. 2 is a flowchart showing the flow of the clustering process, and FIG. 3 is a conceptual diagram for explaining the clustering process in the present embodiment, and each circle indicates a learner. It is assumed that the learners are distributed as shown in FIG. 3A in the feature vector space of the acquired characteristics.
The cluster extraction unit 3 initializes the cluster number k that is the number of the generated cluster in the internal storage unit, that is, sets “k = 0” (step S1).

Next, the average characteristic amount calculation unit 2 calculates the average of the achievement levels for all items of the respondents (learners) based on the input test response information, and an average feature including the average of the achievement levels of the respective items. The amount Vk is output (step S2).
The characteristic amount calculation unit 1 calculates the characteristic amount Vs (i) for each learner i (1 ≦ i ≦ n, n; the number of learners included in the learner group) and is obtained. The feature quantity Vs (i), test answer information, and learner attribute information are stored in the learner information database 5 in association with the learner identification number (step S3).

Next, the cluster extraction unit 3 measures the similarity between the average feature value Vk and the feature value Vs (i) of all learners i in the learner group (step S4).
Here, for example, the cluster extraction unit 3 obtains the similarity by the reciprocal of the distance of the feature vector of each feature quantity, and uses the Euclidean distance, the Mahalanobis distance, the exponential distance, or the like as the distance. As a clustering method, an aggregation method, a two-way (block clustering) method, a K-means method, or the like can be used.
Then, the cluster extraction unit 3 assumes that learners whose similarity to the average feature amount Vk is equal to or greater than a predetermined threshold value belong to the cluster G0, and associates the identification numbers of these learners with the cluster G0 to obtain the cluster characteristics. Store in the database 6 (step S5).

Next, the cluster extraction unit 3 selects any one of the learners having the feature amount Vs (i) whose similarity is less than a predetermined threshold, and sets the learner's feature amount Vs (i) to a new average. The feature amount Vk (hereinafter referred to as the reference feature amount Vk) is used, the cluster number k is incremented by 1 (k + 1), and the cluster G1 is set, and the process proceeds to step S4 (step S6).
In step S6, the cluster extraction unit 3 increments the cluster number k by 1, and then determines whether or not the cluster number k exceeds the number of learners in the learner group. If the number does not exceed (k ≦ n), the process proceeds to step S4. If the cluster number k exceeds the number of learners (k> n), the process proceeds to step S6.

Next, the cluster extraction unit 3 refers to the identification numbers of learners included in each of the clusters G0 to Gn in the cluster characteristic database 6 and merges the clusters Gk including the same learner identification numbers. As a new one cluster, clusters G0 to G5 are output as a result of extracting clusters having a predetermined tendency of the final learner group as shown in FIG. 3B (step S7).
In addition, the cluster attribute extraction unit 4 refers to the learner information database 5 for attribute information common to learners included in each of the clusters G0 to Gn, and compares the attribute information between the clusters. The cluster attribute feature is extracted, and the characteristic amount of each learner is averaged to calculate a cluster acquisition characteristic, which is stored in the cluster characteristic database 6 corresponding to each cluster.
The cluster attribute extraction unit 4 stores the response status (selection rate of options) for each cluster in the cluster characteristic database 6.

That is, in the above-described clustering (steps S1 to S7), an average learner is obtained by calculating an average of the characteristic amounts of each learner in the learner group, and starting clustering using this average value as the average feature amount Vk. Next, the process of sequentially extracting learner clusters having a tendency different from the average is performed.
In addition, as a reason for extracting learner clusters having average feature quantities in the first stage, as an assumption that the learner's characteristic quantities are different, the average learner follows a certain evaluation axis by class. We can expect to learn more.

On the other hand, learners who have common factors such as lack of specific knowledge (in past learning history) tend to be biased toward the correct answer rate of each item of the test with respect to the average amount of learner characteristics. Will be shown.
Then, different clusters are extracted even among learners who have a biased tendency due to the difference in the item having the lack of knowledge and the degree of the lack of knowledge.

As described above, the characteristic amount of each learner can be obtained, for example, as a feature vector in the number-of-items dimension having the achievement level of each item obtained from the test result as an element.
The similarity of the characteristic quantities between learners can be defined as, for example, a correlation value between the feature vectors (in this embodiment, the reciprocal of the feature vector distance).
Thereby, the cluster G0 obtained as the similarity evaluation with the average feature amount Vk can be considered as a category located in the vicinity of the average value of the characteristic amount, and becomes a category indicating an average learning characteristic.

<Affiliation cluster estimation processing>
Next, with reference to FIG. 1 and FIG. 4, an affiliated cluster estimation process for estimating a reaching cluster when the learner transitions in the current learning state will be described. FIG. 4 is a conceptual diagram for explaining this affiliation cluster estimation process, that is, learner A performs tests # 1 to # 4 in time series, and finally belongs to any of clusters C0 to C5. FIG.

For example, since the above-described clustering process shows the acquisition tendency for each subject in the semester, the four tests such as the test # 1 to the test # 4 such as the middle, the end of the term, and the small test in the semester are performed. It is determined by the characteristic amount obtained from the achievement level of each item.
However, for each learner, assuming that the current learning state has changed from the achievement level of each item in each test on the way, if the cluster that will eventually belong in the semester can be estimated, each learning at that time It may be possible to obtain guidance guidelines for learners and guide learners to an average cluster of learners.

Here, the cluster characteristic database 6 includes cluster attribute features of clusters C0 to C5 to which each learner belongs in the learner group of the previous year (a plurality of knowledge confirmation tests that are considered necessary for each subject performed at the start). The evaluation characteristics (feature vector results are also included depending on the item) and cluster acquisition characteristics are recorded in association with each cluster.
Further, the learner A is made to perform the knowledge confirmation test, and the attribute information already described is obtained through a questionnaire or the like, and stored in the learner database 5 corresponding to the identification number of each learner A. .

  Then, at each stage of the tests # 1, # 2, # 3, and # 4, for example, at the end of the test # 1, the affiliation cluster estimation unit 9 inputs the test answer information of the learner A, and the item of the test # 1 And the achievement level of the item corresponding to the item of test # 1 is extracted from the cluster acquisition characteristics of each cluster stored in the cluster characteristic database 6 and compared with the partial characteristic quantity. At the same time, the attribute information of the learner A and the cluster attribute feature of each cluster are compared, the similarity is obtained from these comparisons, and the learner A is the cluster to which the learner A belongs at the current stage.

Here, as described in the clustering process, the similarity is obtained by calculating the distance between each feature vector and calculating the reciprocal of this distance.
Further, when the item used in the input test answer information is an unknown item (an item not used when a cluster stored in the cluster characteristic database 6 is extracted), metadata of the item ( With reference to the learning target), it is possible to perform cluster estimation processing by using the same item for which the metadata matches.
Classes with similar learning content are conducted every year, but can be used effectively when the test items are different.

<Analysis result display function>
With reference to FIG. 5, a case where the analysis process of the learning characteristics of the learner based on the entire test result is displayed at once will be described.
When the user inputs an instruction to display an estimation result of the learner's cluster from the input unit, the learning characteristic display unit 10 displays an input field for an identification number. Here, when “ABC123” is input, the acquisition characteristic display unit 10 displays the graph of R1 indicating the acquisition characteristic of each cluster of the upper cluster, middle cluster, lower cluster and each cluster of the extraction cluster on the display unit. To do.

And the acquisition characteristic display part 10 plots the achievement rate with respect to each item of the individual learner corresponding to the identification number of the learner which the user input in the graph R1.
Here, the learning characteristic display unit 10 sorts and displays the items in the order of the achievement rate, using the specified cluster or the element value of the individual learner as a key, for example, in the order of the achievement rate. It is also possible.
In addition, the cluster attribute feature display unit 11 displays the cluster (here, cluster B) number determined to belong to the learner ABC123 as, for example, the text frame R2 in association with the graph R1, and the cluster B Cluster attribute features such as “low mathematics completion rate”, “low differential performance” and “good English” are displayed.

Next, the item information display unit 14 corresponds to the selected item content, that is, the selected item, as the text frame R3 when the user clicks and selects the item number on the horizontal axis of the graph R1 with a mouse or the like. The question content (question sentence, choice, etc., for example, question 2; answer the following question *****, *****) is displayed.
Then, the item answer status display unit 12 allows the user to select the item number on the horizontal axis of the graph R1 by clicking with the mouse or the like, and for the selected item, the upper cluster, the middle cluster, the lower cluster, and the learner. The selection ratio of the options in each of the whole is displayed on the graph R4.

Here, the choices of item 2 include A, B, C, and D. For example, the ratio of learners who have selected option A is 19.47% in the upper cluster and 19.78% in the middle cluster. Yes, the lower cluster is 21.24, and the overall average is 20.1%.
If the selected item is not an option as described above, it is possible to perform processing equivalent to the option by classifying the answers manually or by performing the same determination by text matching processing.

In addition, the item achievement status display unit 13 displays image information as shown in the graph R5 that allows visualization of the learner's achievement status in consideration of item characteristics.
That is, the item achievement status display unit 13 uses the item difficulty (item achievement, correct answer rate, difficulty, etc.) and item discrimination (point biserial correlation coefficient, etc.) as item characteristics, Each item is arranged on two horizontal axes.

In the graph R5, the relationship between the difficulty level and the discrimination force for the items included in the test is shown, the vertical axis indicates the discrimination force, and the horizontal axis indicates the difficulty level.
Furthermore, the item achievement status display unit 13 displays the item achievement status of the item of the designated respondent (learner) on the graph R5.
At this time, the item achievement status display unit 13 displays the item achievement status of the designated learner in different colors depending on, for example, the correct answer, the incorrect answer, or the degree of achievement.

Next, with reference to FIG. 6, the case where the estimation process of the learner's affiliation cluster based on the test result in the middle is performed and the estimation result is displayed will be described.
FIG. 6 shows a conceptual diagram in which the result of clustering processing (clustering of each of the cluster extraction unit 3 and the item characteristic calculation unit 7) is displayed on a display unit (not shown). In FIG. 6, as in FIG. 5, the horizontal axis indicates the number of each item of the test, and the vertical axis indicates the achievement rate of each item (unless it is shown in FIG. 5, the test items until midway).
The learning characteristic display unit 10 displays an input field for an identification number by the user performing control to display an estimation result of the learner's cluster from the input unit, and the individual identification number of the learner is displayed with respect to the input field. Here, when “ABC567” is input, the learning characteristic display unit 10 displays a graph of Z1 indicating the cluster learning characteristic of each extracted cluster on the display unit.

And the acquisition characteristic display part 10 plots the achievement rate with respect to each item of the individual learner corresponding to the identification number of the learner which the user input in the graph Z1 which is the cluster acquisition characteristic.
Here, the learning characteristic display unit 10 sorts and displays the items in the order of the achievement rate, using the specified cluster or the element value of the individual learner as a key, for example, in the order of the achievement rate. It is also possible.
In addition, the cluster attribute feature display unit 11 displays the cluster number (here, cluster B) that the learner ABC567 is estimated to belong to, for example, as the text frame Z2, corresponding to the graph Z1 that is the cluster acquisition characteristic. The cluster attribute characteristics of this cluster B, for example, “low math completion rate”, “low differential performance”, “good English” and “low achievement rate of learning target α” are displayed.

Next, the item information display unit 14 corresponds to the selected item content, that is, the selected item as the text frame Z3 when the user clicks and selects an item number on the horizontal axis of the graph Z1 with a mouse or the like. The question content (question sentence, choice, etc., for example, question 10; answer the following question ***) and the metadata of the question (learning target β: obtain area of polygon, etc.) are displayed.
Here, it is assumed that the metadata has a creation date of the item, a creator, and a learning target by which achievement confirmation is performed by default, and the display of metadata required by the user can be designated.
According to the above-mentioned process, by inputting test response information up to the middle of the semester, it is estimated which cluster the individual belongs to at present (which cluster will belong to the current learning situation) It can be used as a guideline for correction instruction of a corresponding learner in a class of a subject whose learning content is planned.

As described above, according to the learner acquisition characteristic analysis system of the present invention, the learning tendency of each individual learner is obtained by each display process, such as a plurality of acquisition characteristics, a positional relationship with the cluster to which the learner belongs, and item characteristics. It becomes possible to observe from the viewpoint of evaluation.
As a result, the learner acquisition characteristic analysis system of the present invention can efficiently grasp the tendency of each learner to achieve his / her learning goals, how to make mistakes, lack of prerequisite knowledge, etc. based on cluster attribute characteristics. Effective learning measures such as prioritizing items to be strengthened can be made.

Further, according to the learner acquisition characteristic analysis system of the present invention, in the display of the graph R1 and the graph Z1 of the acquisition characteristic display unit 10, by displaying the characteristics of the cluster and each individual learner at the same time, It is possible to easily compare the learning tendency with its own learning tendency, and it is possible to grasp its own learning tendency in the collection end vehicle group.
Furthermore, according to the learner acquisition characteristic analysis system of the present invention, it is possible to easily grasp visually each of the clusters and the items that the learner is not good at by sorting and displaying the designated keys.
At this time, if a key (for example, individual learner) is specified and sorted, the distance from each broken line becomes wide, so that the difference from other than the key (each cluster) can be easily understood visually.

Further, according to the learner acquisition characteristic analysis system of the present invention, by displaying the item achievement status, the item characteristic of the correct / incorrect answer item is visualized together with the individual achievement level of the learner, so that the learner It is possible to visually grasp whether a score is obtained by achievement of various items.
At this time, by checking in conjunction with the display images (display information) of the graph R1 and the graph Z1 of the acquired characteristic display unit 10, it is possible to efficiently grasp the characteristics of the item.

In the display of the graph R5 in FIG.
-On the horizontal axis of the graph, items that are displayed to the left are easier to answer correctly, while items that are displayed to the right are more difficult to answer correctly. Items that cannot be answered correctly, on the other hand, items that are displayed closer to the bottom are items that have a smaller relationship between the achievement level of the items and the results.

Furthermore, according to the learner acquisition characteristic analysis system of the present invention, it is possible to know the tendency of individual mistakes of the learner by displaying the item response status, that is, to learn what kind of cluster the behavior is similar to. The person himself can know.
In addition, according to the learner acquisition characteristic analysis system of the present invention, it is possible to confirm what kind of item it is by displaying the item content.

  Note that a program for realizing the function of the learner acquisition characteristic analysis system in FIG. 1 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system and executed. The above-described learner acquisition characteristic analysis process may be performed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer system” includes a WWW system provided with a homepage providing environment (or display environment). The “computer-readable recording medium” refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, and a CD-ROM, and a storage device such as a hard disk built in the computer system. Further, the “computer-readable recording medium” refers to a volatile memory (RAM) in a computer system that becomes a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. In addition, those holding programs for a certain period of time are also included.

  The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

It is a block diagram which shows the structural example of the learner acquisition characteristic analysis system by one Embodiment of this invention. It is a flowchart which shows the operation example of the process of the clustering in the learner acquisition characteristic analysis system of FIG. It is a conceptual diagram for demonstrating the clustering process of the flowchart shown in FIG. It is a conceptual diagram for demonstrating operation | movement of the cluster estimation in the learner acquisition characteristic analysis system of FIG. It is explanatory drawing of the display screen explaining the process which displays the result of the analysis process of the learner's acquisition characteristic by the whole test result at once. It is explanatory drawing of the process which displays the result of the estimation process of the cluster to which a learner belongs in the middle of a semester on a display screen.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Characteristic amount calculation part 2 ... Average characteristic amount calculation part 3 ... Cluster extraction part 4 ... Cluster attribute extraction part 5 ... Learner information database 6 ... Cluster characteristic database 7 ... Item characteristic calculation part 8 ... Item information database 9 ... Affiliation cluster Estimating unit 10 ... acquisition characteristic display unit 11 ... cluster attribute feature display unit 12 ... item answer status display unit 13 ... item achievement status display unit 14 ... item information display unit

Claims (11)

A characteristic amount calculation unit that outputs the achievement level of each item unit for each learner as a characteristic amount;
An average characteristic amount calculation unit that calculates an average value of achievement for each item and outputs as an average characteristic amount;
A learner acquisition characteristic analysis system comprising: a cluster extraction unit that performs clustering of learners in a predetermined learner group based on the average characteristic quantity and the characteristic quantity. The cluster extraction unit
The learning according to claim 1, wherein at the start of clustering, the average characteristic amount is compared with the characteristic amounts of all learners, and learners having characteristic amounts equal to or higher than a predetermined similarity are output as average clusters. Acquisition characteristics analysis system. A database that stores attribute information including past learning history of each learner, grades of subjects taken, strengths and weaknesses, and subjects of interest;
The learner acquisition characteristic according to claim 1 or 2, further comprising: a cluster attribute extraction unit that extracts attribute information common to learners belonging to a cluster from the database for each extracted cluster. Analysis system.   When multiple stages of tests are performed in time series during a given period and clustering is performed based on the results of the multiple stages, the test learner's test characteristic amount to be estimated and the previously created cluster When the target learner learns in the current state by comparing the test learner's characteristic quantities at the same stage and extracting a cluster containing comparative learners with high similarity to the target learner The learner acquisition characteristic analysis system according to any one of claims 1 to 3, further comprising a belonging cluster estimation unit that estimates a cluster included in the future.   For each cluster extracted by the degree of similarity, the belonging cluster estimation unit compares the attribute information common to each comparison learner belonging to the cluster with the attribute information possessed by the estimation learner, and the cluster having a high degree of matching of the attribute information The learner acquisition characteristic analysis system according to claim 4, wherein: A characteristic amount calculation process in which the achievement level of each item in each learner is output as a characteristic amount,
An average characteristic amount calculation process in which the average value of achievement for each item is calculated and output as an average characteristic amount,
A learner acquisition characteristic analysis method comprising: a cluster extraction process in which each learner in a predetermined learner group is clustered based on the average characteristic quantity and the characteristic quantity. In the cluster extraction process,
The average characteristic amount is compared with the characteristic amounts of all learners at the start of clustering, and learners having characteristic amounts equal to or higher than a predetermined similarity are output as average clusters. Learner acquisition characteristics analysis method. A memory process for storing attribute information including past learning history of each learner, grades of subjects taken, weak subjects, and subjects of interest in a database,
The learner learning according to claim 6 or 7, further comprising: a cluster attribute extraction process in which attribute information common to learners belonging to a cluster is extracted from the database for each extracted cluster. Characteristic analysis method.   When multiple stages of tests are performed in time series during a given period and clustering is performed based on the results of the multiple stages, the test learner's test characteristic amount to be estimated and the previously created cluster When the target learner learns in the current state by comparing the test learner's characteristic quantities at the same stage and extracting a cluster containing comparative learners with high similarity to the target learner The learner acquisition characteristic analysis method according to claim 6, further comprising a belonging cluster estimation process in which a cluster included in the future is estimated. A characteristic amount calculation process in which the achievement level of each item unit for each learner is output as a characteristic amount;
An average characteristic amount calculation process in which an average value of achievement levels for each item is calculated and output as an average characteristic amount;
A computer-executable program for performing a learner acquisition characteristic analysis process including: a cluster extraction process in which each learner is clustered based on the average characteristic quantity and the characteristic quantity. In the cluster extraction process,
The learning according to claim 10, further comprising a process in which the average characteristic amount is compared with characteristic values of all learners at a clustering start time, and a learner having a characteristic amount equal to or higher than a predetermined similarity is output as an average cluster. A computer-executable program that performs a person-acquired characteristic analysis process.

Images