JP6566468B2 - User-adaptive test program, apparatus and method for identifying answers based on ambiguous understanding and minor misunderstandings - Google Patents

Description

  The present invention relates to a technique of an e-Learning system that determines the academic ability of a user (learner).

  According to the user-adaptive e-Learning system, using a database that accumulates the answer trends of a large number of users in the past as a history, for each user, based on the answer results, You can select and specify content. There is also a “testing” system that provides different problems and learning materials for each user by placing the e-Learning system as a server on the network.

  It is very difficult to create a plurality of problems corresponding to each user's level of understanding. In particular, it is extremely difficult to make the difficulty level constant for each question and determine the level of understanding. On the other hand, there is an adaptive testing (computer adaptive test) technique that estimates the parameter of a question to be given each time and selects the next problem according to the answer result of the user (for example, non-patent literature). 1).

  There is also a technique for determining the degree of understanding of a user from a test database and test execution results (see, for example, Patent Document 1, Non-Patent Documents 2 and 3). According to this technology, it is possible to select a question according to the user's comprehension probability and give a question, and to determine the user's academic ability with high accuracy even with a small number of problems. In particular, according to the technique described in Non-Patent Document 3, the problem of maximizing the network-type information amount for the learning model is controlled by using a Bayesian network.

JP 2004-170842 A

Masaomi Ueno, “Evaluation of Learning in the New Era”, Education Test Research Center 18th Research Group Report, June 18, 2010, [online], [Search April 21, 2015], Internet <https: / /www.cret.or.jp/files/dbd90c4cb469bf94ef5da2256616c7e7.pdf> Masaomi Ueno, Hitoshi Onishi, Kazuo Shigeki, “Proposal of Test Theory with Probability Network”, IEICE Technical Report, ET93-64, 55-62,1993, [online], [Search April 21, 2015 ] Internet <http://ci.nii.ac.jp/naid/110003192386> Masaomi Ueno, Chiharu Kawatake, Hitoshi Onishi, “Structural Optimization Method in Network Type Test Theory”, IEICE Technical Report, ET93-ET137, March 26, 1994, IEICE, Vol100, P50-P60, [online ], [Search March 8, 2015], Internet <URL: http://jglobal.jst.go.jp/public/20090422/200902114032577390> Kazuo Hangeki, Masaomi Ueno, Yoichi Motomura, “Outline of Bayesian Network”, Baifukan, July 21, 2006

  However, even if the user himself / herself does not understand the problem, if the answer is accidentally correct (ambiguous understanding), or if the problem is completely understood, the user may inadvertently make a wrong answer (minor misunderstanding). ). For example, it may change depending on the environment (location and time) for the user and the psychological state of the user.

  On the other hand, according to the conventional testing system, an attempt is made to determine the user's academic ability as accurately as possible by increasing the number of questions to be presented to the user and diversifying the variations of the problems. Even if the answer does not match the user's understanding and the answer result, it is attempted to be solved by increasing the number of problems, using it as a category of error. In other words, even if the academic achievement measurement has the possibility of accidental coincidence or inconsistency, it operates on the assumption that the user's answer always corresponds to the user's understanding.

  On the other hand, can the inventor of the present application identify an answer based on an ambiguous understanding or a slight misunderstanding? I thought. In order to estimate the degree of understanding when there is an answer based on an ambiguous understanding or minor misunderstanding, the answer based on the ambiguous understanding or minor misunderstanding should be carefully identified as a “judgment note”. At this time, we thought that the degree of understanding of the original user should be estimated by giving further questions related to the problem. Also, by carefully discriminating only the learning units that have been determined to be judged by the user, it leads to a more accurate estimation of the user's understanding, resulting in a reduction in the number of questions for determining the user's academic ability May lead to I thought.

  Accordingly, an object of the present invention is to provide a user-adaptive test program, apparatus, and method that can identify an answer based on an ambiguous understanding or a slight misunderstanding.

According to the present invention, a user-adaptive test program for causing a computer mounted on an apparatus to function is provided.
Learning unit storage means that stores in advance a directed graph in which learning units including questions and answers are connected as nodes, and learning units are connected with arcs of prior understanding probabilities;
Understanding degree probability update means for calculating the subsequent degree of understanding probability in all other learning units that have not been asked for correct / incorrect answers of questions included in the already given learning unit for each user;
For each user, a problem presentation means for presenting a problem included in an unquestioned learning unit,
Answer determination means for determining the correct / incorrect answer to the question for each user;
When a correct answer is given to a question of a learning unit whose ex-post understanding probability is lower than the determination threshold, or when a wrong answer is given to a question of a learning unit whose understanding probability is higher than the determination threshold, the learning unit Let the computer function as a judgment caution identification means to identify a judgment caution as understanding / misunderstanding is unknown ,
The problem presenting means displays the problem included in the learning unit having the largest change in the understanding probability before and after recalculating the understanding probability of all other learning units that have not been asked by the understanding probability update means. To answer questions
Characterized Rukoto cause the computer to function as.

According to another embodiment of the test program of the present invention,
The directed graph of the learning unit storage means preferably has the computer function so that a plurality of learning units are nodes and arcs are connected by a parent-child relationship according to the learning time and learning content similarity.

According to another embodiment of the test program of the present invention,
The comprehension probability update means is a Bayesian network, and each time a correct answer / wrong answer of the learning unit determined by the answer determination means is output, if the answer is a correct answer, the understanding probability of the learning unit It is also preferable that the computer understands the understanding level probability = 0 of the learning unit in the case of an incorrect answer, and causes the computer to function to recalculate the understanding level probabilities of all other unlearned learning units.

According to the present invention, an apparatus for performing a user adaptive test comprising:
Learning unit storage means that stores in advance a directed graph in which learning units including questions and answers are connected as nodes, and learning units are connected with arcs of prior understanding probabilities;
Understanding degree probability update means for calculating the subsequent degree of understanding probability in all other learning units that have not been asked for correct / incorrect answers of questions included in the already given learning unit for each user;
For each user, a problem presentation means for presenting a problem included in an unquestioned learning unit,
Answer determination means for determining the correct / incorrect answer to the question for each user;
When a correct answer is given to a question of a learning unit whose ex-post understanding probability is lower than the determination threshold, or when a wrong answer is given to a question of a learning unit whose understanding probability is higher than the determination threshold, the learning unit Understanding / misunderstanding possess the identifying determination note identification means and the determination noted as being unknown,
The problem presenting means displays the problem included in the learning unit having the largest change in the understanding probability before and after recalculating the understanding probability of all other learning units that have not been asked by the understanding probability update means. It is characterized by the questions .

According to the present invention, there is provided a test method for a device for performing a user adaptive test,
The apparatus has a learning unit storage unit that stores in advance a directed graph in which learning units including questions and answers are connected as nodes, and learning units are connected with arcs of prior understanding probabilities,
The device
For each user, a first step of calculating a subsequent understanding probability in all other learning units that have not been answered for correct / incorrect questions included in the existing learning unit;
For each user, a second step of presenting questions included in an unquestioned learning unit;
A third step for each user to determine the correct / incorrect answer to the question;
When a correct answer is given to a question of a learning unit whose ex-post understanding probability is lower than the determination threshold, or when a wrong answer is given to a question of a learning unit whose understanding probability is higher than the determination threshold, the learning unit A fourth step of presuming a judgment attention as understanding / misunderstanding is unknown;
Run
In the second step, before and after recalculating the understanding probabilities of all other learning units that have not been asked in the first step, the problems included in the learning unit having the largest change in the understanding probability are To answer questions
It is characterized by performing as follows.

  According to the user-adaptive test program, apparatus and method of the present invention, an answer based on an ambiguous understanding or a slight misunderstanding can be identified.

It is a functional block diagram of the test apparatus in this invention. It is a basic flowchart of the present invention. It is an example of the directed graph memorize | stored in the learning unit memory | storage part. It is the 1st explanatory view showing the directed graph and comprehension probability of a learning unit at the time of initial setting. It is the 2nd explanatory view showing the update of the comprehension degree probability of the learning unit when answering the learning unit 5 correctly. It is a 3rd explanatory drawing showing the update of the understanding degree probability of the learning unit in the case of having made a wrong answer to the learning unit 8. It is the 4th explanatory view showing renewal of the comprehension degree probability of a learning unit in case of having answered wrong to learning unit 4. It is the 5th explanatory view showing the update of the comprehension degree probability of the learning unit when answering the learning unit 13 correctly.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a functional configuration diagram of a test apparatus according to the present invention.
FIG. 2 is a flowchart of the present invention.

  According to FIG. 1, a test apparatus 1 that performs user-adaptive testing and a terminal 2 (for example, a tablet terminal) operated by a user to be learned are connected via a network. The test apparatus 1 functions as a server for the terminal 2. Further, the functional configuration unit of the test apparatus 1 may be mounted on the terminal 2 so that the terminal alone functions as the test apparatus.

  1 includes a learning unit storage unit 10, an understanding degree probability update unit 11, a problem presentation unit 12, an answer determination unit 13, and a determination attention identification unit 14. These functional components are realized by executing a program that causes a computer mounted on the apparatus to function. Further, the processing flow of these functional components can be understood as a user-adaptive test method as shown in FIG.

[Learning unit storage unit 10]
The learning unit storage unit 10 stores in advance a directed graph in which learning units are nodes and the learning units are connected with arcs of prior understanding probabilities.

  FIG. 3 is an example of a directed graph stored in the learning unit storage unit.

  The list (directed graph) in FIG. 3 is obtained by connecting arcs by a parent-child relationship according to the learning time and learning content similarity with a plurality of learning units as nodes. Here, the understanding probability of the previous learning unit affects the understanding probability of the subsequent learning unit according to the learning content similarity. As the “learning time”, for example, it may be possible to roughly specify the relationship from the front to the back according to the learning schedule, such as April-> May of the first grader of junior high school. Further, the “learning unit” includes a question and an answer, and may use not only text but also an image, video, and sound.

  The “learning content similarity” between learning units may be a comparison between word sets obtained by extracting texts of learning units by morphological analysis. Morphological analysis is a process that uses a grammar and word dictionary as an information source to divide a sentence written in a natural language into morphemes (Morpheme), which are the smallest units that have meaning as a language, and discriminate each part of speech. . Here, all the extracted words may be used, or only morphemes whose part of speech is a noun may be extracted.

  As a method for calculating the learning content similarity, for example, a TF-IDF (Term Frequency-Inverse Document Frequency) method may be used. According to this method, TFi, j calculated from the number and frequency of occurrence of the word wi in the learning content of the learning unit j and the learning unit in which the word wi appears among all the learning units j = 1 to N. Using the IDFi from which the number mi is calculated, a weight for the word wi included in the learning unit j is used. For learning units, the learning content is expressed in Bag-of-Words, and the components of each vector are weighted by the corresponding value of TF × IDFi. Bag-of-Words expresses learning text as a feature vector defined by the frequency of one word, and derives similarity between sentences using IDFs derived in advance based on sentence sets as word weights. is there. Then, a cosine distance is calculated between the two vectors, and this is used as the similarity between the learning contents.

  Learning units whose similarity is equal to or greater than a predetermined threshold are connected by arcs of a directed graph. Based on the similarity of the learning unit, for example, it is considered that the understanding probability in the learning unit after the user is connected by the arc changes corresponding to the understanding probability in the previous learning unit.

  This directed graph may be created manually by considering the relevance between learning units based on the learning guidance policy. Moreover, it may be automatically created based on the learning start time from the similarity of the learning content in each learning unit. For example, if the dependency of each learning unit in the same subject is clear, it can be determined relatively easily. According to FIG. 3, the single subject “mathematics” is applied. However, the subject may extend over a plurality of subjects, or may be based on highly specialized qualification education or corporate education.

[Understanding Level Probability Update Unit 11 (S11)]
The comprehension degree probability update unit 11 calculates, for each user (learner), the comprehension degree probabilities of all other unlearned learning units for correct / incorrect answers to questions included in the already presented learning unit. . Specifically, the user sequentially answers the questions to be asked, and collates the correct / incorrect answer results with an understanding level estimation model (directed graph). Then, the correct answer rate (understanding probability) for unquestioned questions is updated for each user.

  It is assumed that the understanding level probability updating unit 11 calculates the understanding level probability for each learning unit in a comprehensive manner for all learning units for a large number of users as an initial state. For example, the understanding probability in the initial state may be an average correct answer rate in each learning unit of a large number of users. Of course, it may be preset manually or may be calculated using machine learning.

  As the understanding level estimation model of the understanding level probability update unit 11, for example, a “Bayesian network” is used. Every time the correct / incorrect answer of the learning unit determined by the answer determination unit 13 to be described later is output, the comprehension probability (likelihood) is recalculated for all other learning units not yet presented.

  Specifically, the understanding level probability update unit 11 stores, as an answer, the understanding level probability of the learning unit in the case of a correct answer = 1.00, and in the case of an incorrect answer, the understanding level probability of the learning unit = 0.00. Recalculate the comprehension probabilities of all other learning units that have not been asked. As a result, it is possible to update the comprehension probability of an unquestioned question using the answer result of the question that has already been set.

  A Bayesian network is a graphical network that describes a causal relationship between nodes by probability (see, for example, Non-Patent Document 4). The network in the present invention refers to a graph structure in which the learning unit is “node” and the similarity given to arcs between nodes is given as “weight”. A Bayesian network is a model of probabilistic reasoning that expresses inference of causal relations between a plurality of nodes by a directed acyclic graph structure and expresses a relation between individual variables by a conditional probability. A directed acyclic graph structure is one in which an arrow is attached to an arc (directed), and the path of the arrow does not circulate through a node. By using a Bayesian network, the correct answer rate (understandability probability) of uncertain events (correct answer / incorrect answer) for the nodes on the route that are given after the correct answer / incorrect answer of each node already answered. It can be estimated quantitatively.

For a Bayesian network, for example, consider a case of parent node A-> child node B, parent node A-> child node C as a directed graph between nodes A, B, and C.
A comprehension probability: P (A)
B comprehension probability: P (B)
C comprehension probability: P (C)
C comprehension probability when A answers correctly = P (C | A)
C comprehension probability when B answers correctly = P (C | B)
C comprehension probability when A answers correctly and then B answers correctly
= P (C | A) P (C | B)
Every time an answer result of a question that has been given is obtained, the understanding probabilities of all the unanswered nodes (learning units) are recalculated. As a result, even in a complex system, it is possible to model a probabilistic dependency relationship in a problem to be presented subsequently using the understanding probability (conditional probability) at each node.

  The directed graph of the present invention is a network that specifies a parent-child relationship between learning units. Therefore, when a correct answer is given for an arbitrary learning unit, the comprehension probability of the learning unit connected to the learning unit is also increased according to the correct answer unit. On the other hand, when an incorrect answer is made for an arbitrary learning unit, the comprehension probability of the learning unit connected to the learning unit is also lowered according to the erroneous answer unit.

[Problem Presentation Unit 12 (S12)]
The problem presentation unit 12 presents a problem included in an unquestioned learning unit for each user. The learning unit to be set next is selected by one of the following four methods, for example.

(First learning unit selection method)
The problem presenting unit 12 then questions the problem included in the learning unit having the understanding level probability closest to the determination threshold. For example, a problem with an understanding probability close to the determination threshold of “I understand” or “I don't understand” is more important for the user's learning determination. That is, it can be said that the problem closer to the determination threshold has the highest amount of information for all problems. For such a problem, the answer is updated to comprehension probability = 1.00 or 0.00, which affects the update of the comprehension probability for other unlearned learning units.

(Second learning unit selection method)
The problem presenting unit 12 next issues a problem included in the learning unit that becomes a parent node for the learning unit identified as the determination attention by the determination attention identifying unit 14. The learning unit is composed of a directed graph. For this reason, for example, a teacher can review a learning unit of “expression using characters” and “spatial figure” as a parent node of a learning unit for a student who does not sufficiently understand “primary function”. . Students can re-learn by returning to the previous learning unit and reviewing. This ultimately allows the user to accumulate systematic understanding.

(Third learning unit selection method)
The problem presenting unit 12 re-calculates the understanding level probabilities of all other unlearned learning units by the understanding level probability updating unit 11 before and after the problem included in the learning unit with the largest change amount of the understanding level probability. Next, we will ask questions. The user's answer to one question will update the comprehension probability of all other learning units that have not been asked. The larger the change in comprehension probability, the more ambiguous the user's answer is. It is thought to include understanding and minor misunderstandings. Such learning units should pay attention to estimation of comprehension.

(Fourth learning unit selection method)
When the learning unit includes a plurality of questions and answers, the problem presenting unit 12 performs the following problem other than the question on the learning unit identified as the determination attention by the determination attention identification unit 14 described later. Ask questions. That is, when a plurality of questions of the same type are given and the correct answer rate of these questions is equal to or higher than a predetermined ratio, the learning unit can be updated as understood (understandability probability = 1.00). For example, an average of 80 points (predetermined ratio) or more can be determined as “understanding” by retesting five questions of the same learning unit that have been determined to be determined. That is, when the correct answer is 4 or more out of 5 questions, the comprehension probability is updated to 1.00, and when the correct answer is 2 questions, the comprehension probability is updated to 0.00.

(Learning unit selection method based on another embodiment)
By combining the above-described second to fourth learning unit selection methods, the problem presentation unit 12 can also select a problem to be applied next.
When the user whose understanding level probability is equal to or less than the first predetermined threshold correctly answers, the “second learning unit selection method” is selected. Specifically, the problem included in the learning unit that becomes the parent node with respect to the learning unit identified as the determination attention by the determination attention identifying unit 14 is given next.
If the user whose understanding probability is greater than or equal to the second predetermined threshold (> first predetermined threshold) gives an incorrect answer, the “fourth learning unit selection method” is selected. Specifically, the learning unit includes a plurality of questions and answers, and for the learning unit identified as the determination attention by the determination attention identifying unit 14, questions other than the question that has been presented are next presented.
The “third learning unit selection method” is selected for the user whose comprehension probability is other than those described above. Specifically, the problem included in the learning unit having the largest change in the understanding probability before and after the understanding probability update unit 11 recalculates the understanding probability of all other learning units that have not been asked. Next is the question.

  Then, the problem presentation unit 12 clearly indicates the selected problem to the user. When the test apparatus 1 is a server, the problem presentation unit 12 transmits the problem content to the terminal 2 operated by the user.

[Answer determination unit 13 (S13)]
The answer determination unit 13 determines the correct / incorrect answer to the question for each user. Since the learning unit includes the question and the answer, the correct answer / wrong answer is determined by comparing the answer of the user with the answer of the question. The determination result is output to the determination attention identifying unit 14.

[Determination Attention Identification Unit 14 (S14)]
When the determination attention identifying unit 14 correctly answers a question of a learning unit whose understanding probability is lower than the determination threshold, or when it incorrectly answers a question of a learning unit whose understanding probability is higher than the determination threshold, It is identified as “judgment attention” because the understanding / misunderstanding of the learning unit is unknown.
Correct answer to a problem with low probability of understanding-> Accidentally correct answer
There is a possibility of vague understanding (the lower the probability of understanding (closer to 0.0), the higher the probability of not understanding)
Answered a question with a high probability of understanding-> Accidental answer
There is a possibility of minor misunderstanding (the higher the degree of understanding probability (close to 1.0), the higher the possibility of understanding)
Basically, the judgment attention identifying unit 14 understands the understanding degree probabilities obtained from the understanding degree estimation model (the directed graph of the understanding degree probabilities for each learning unit updated by the understanding degree updating unit 11), and correct / incorrect answers as question results. The likelihood of the result is evaluated by comparing the answer.

  Here, the “determination threshold value” of the understanding probability can be arbitrarily set. For example, when the understanding criterion is “I understand” and “I don't understand”, the determination threshold can be set to 0.5. Also, the determination threshold can be changed according to the target for each user or according to the difficulty level of the test problem.

  4 to 8 show a process in which the understanding unit probability of the learning unit is updated according to the present invention.

  FIG. 4 is a first explanatory diagram showing a directed graph and understanding probability of a learning unit at the time of initial setting.

  According to FIG. 4, the understanding probability of each learning unit at the initial setting may be an average value obtained from the answer results of past correct / incorrect answers of many users. The directed graph represents the progress of general understanding of a learning unit and the average understanding probability of each learning unit. The determination threshold is set to 0.5, and the learning unit closer to the determination threshold is relatively difficult to determine “understand” or “not understand”.

  First, the learning unit 5 (= 0.52) closest to the determination threshold = 0.5 is selected, and the problems included in the learning unit are presented to the user. Here, since the learning unit 5 has the understanding probability = 0.52, it is estimated that the learning unit 5 is “understanding” at first.

  FIG. 5 is a second explanatory diagram showing the update of the understanding degree probability of the learning unit when the correct answer is given to the learning unit 5.

  According to FIG. 5, the user has answered the learning unit 5 correctly. At this time, since the “correct answer” is given to the learning unit 5 that has been estimated to be “understood”, it is not estimated that the user has accidentally answered correctly. The understanding unit probability of the learning unit 5 is updated to 1.00, and all other ununderstood understanding probabilities are recalculated by the Bayesian network.

  Next, the learning unit 8 (= 0.47) closest to the determination threshold = 0.5 is selected, and the problems included in the learning unit are presented to the user. Here, since the learning unit 8 has an understanding probability = 0.47, it is initially estimated that the learning unit 8 does not understand.

  FIG. 6 is a third explanatory diagram showing the update of the understanding degree probability of the learning unit when the learning unit 8 is answered incorrectly.

  According to FIG. 6, the user has erroneously answered the learning unit 8. At this time, it is not estimated that the user accidentally made a mistake because the learning unit 8 that was presumed to be “not understood” was “wrongly answered”. The understanding unit probability of the learning unit 8 is updated to 0.00, and all other understanding degree probabilities of unanswered questions are recalculated by the Bayesian network.

  Next, the learning unit 4 (= 0.54) closest to the determination threshold = 0.5 is selected, and the problems included in the learning unit are presented to the user. Here, since the learning unit 4 has an understanding probability = 0.54, it is estimated that the learning unit 4 “understands” at first.

  FIG. 7 is a fourth explanatory diagram showing the update of the understanding degree probability of the learning unit when the learning unit 4 is answered incorrectly.

  According to FIG. 7, the user has erroneously answered the learning unit 4. At this time, since the learning unit 4 that has been estimated to be “understood” has been “incorrectly answered”, it is determined that the user has accidentally answered “notice”.

  Originally, since the answer was incorrect, the understanding unit probability of learning unit 4 is updated to 0.00. However, since the learning unit 4 is identified as the determination attention, the problem presentation unit 12 carefully determines the degree of understanding of the user by any one of the four methods described above. If it is finally determined that the user understands, the learning unit 4 is updated to an understanding probability of 1.00 = 1.00, and all other understanding probabilities of the unanswered questions are recalculated by the Bayesian network.

  Next, the learning unit 13 (= 0.45) closest to the determination threshold = 0.5 is selected, and questions included in the learning unit are presented to the user. Here, since the learning unit 13 has the understanding probability = 0.45, it is estimated that the learning unit 13 does not understand at first.

  FIG. 8 is a fifth explanatory diagram showing the update of the understanding degree probability of the learning unit when the correct answer is given to the learning unit 13.

  According to FIG. 8, the user has correctly answered the learning unit 13. At this time, since the “correct answer” is given to the learning unit 13 that has been estimated to be “not understood”, it is determined as “judgment attention” if the user accidentally answers correctly.

  Since the answer is correct, the learning unit 13 is updated to the understanding probability = 1.00. However, since the learning unit 13 is identified as the determination attention, the problem presentation unit 12 carefully determines the degree of understanding of the user by one of the four methods described above. When it is finally determined that the user does not understand, the learning unit 13 updates the understanding unit probability to 0.00, and the Bayesian network recalculates all other ununderstood understanding probabilities.

  By repeating the processes shown in FIGS. 4 to 8, a learning unit having an ambiguous understanding or a slight misunderstanding is estimated, and the user's level of understanding is determined as carefully as possible for such a learning unit.

  As described above in detail, according to the user-adaptive test program, apparatus, and method of the present invention, it can be applied to an e-Learning system to identify an answer based on an ambiguous understanding or minor misunderstanding. . By increasing the number of questions only for the learning unit identified as the determination attention, the user's degree of understanding can be carefully determined. As a result, in order to finally determine the user's level of understanding compared to the conventional method that assumes that the user's brute force questions and the user's answers always correspond to the user's understanding. The number of questions to be answered can be reduced, and further, the determination error of the comprehension level caused by ambiguous understanding or slight misunderstanding can be reduced.

  Various changes, modifications, and omissions of the above-described various embodiments of the present invention can be easily made by those skilled in the art. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

DESCRIPTION OF SYMBOLS 1 Test apparatus 10 Learning unit memory | storage part 11 Comprehension probability update part 12 Problem presentation part 13 Answer determination part 14 Judgment attention identification part 2 Terminal

Claims (5)

A user-adaptive test program for causing a computer mounted on the apparatus to function,
Learning unit storage means that stores in advance a directed graph in which learning units including questions and answers are connected as nodes, and learning units are connected with arcs of prior understanding probabilities;
Understanding degree probability update means for calculating the subsequent degree of understanding probability in all other learning units that have not been asked for correct / incorrect answers of questions included in the already given learning unit for each user;
For each user, a problem presentation means for presenting a problem included in an unquestioned learning unit,
For each user, an answer determination means for determining the correct / incorrect answer of the answer to the question;
When a correct answer is given to a question of a learning unit whose ex-post understanding probability is lower than the determination threshold, or when a wrong answer is given to a question of a learning unit whose understanding probability is higher than the determination threshold, the learning unit Let the computer function as a judgment caution identification means to identify a judgment caution as understanding / misunderstanding is unknown ,
The problem presenting means includes a problem included in the learning unit having the largest change amount of the understanding degree probability before and after recalculating the understanding degree probabilities of all other learning units not yet asked by the understanding degree probability updating means. Next question
Test program characterized Rukoto cause the computer to function as. The directed graph of the learning unit storage means causes a computer to function by connecting a plurality of learning units as nodes and connecting arcs by a parent-child relationship according to learning time and learning content similarity. Item 2. The test program according to item 1. The comprehension degree probability update means is a Bayesian network, and each time the correct / incorrect answer of the learning unit determined by the answer determination means is output, when the answer is correct, the learning unit It is assumed that the understanding probability is set to 1, and in the case of an incorrect answer, the learning unit is stored as understanding probability = 0, and the computer is made to function so as to recalculate the understanding probabilities of all other learning units that have not been asked. The test program according to claim 1 or 2, characterized in that A device for performing user-adaptive tests,
Learning unit storage means that stores in advance a directed graph in which learning units including questions and answers are connected as nodes, and learning units are connected with arcs of prior understanding probabilities;
Understanding degree probability update means for calculating the subsequent degree of understanding probability in all other learning units that have not been asked for correct / incorrect answers of questions included in the already given learning unit for each user;
For each user, a problem presentation means for presenting a problem included in an unquestioned learning unit,
For each user, an answer determination means for determining the correct / incorrect answer of the answer to the question;
When a correct answer is given to a question of a learning unit whose ex-post understanding probability is lower than the determination threshold, or when a wrong answer is given to a question of a learning unit whose understanding probability is higher than the determination threshold, the learning unit Understanding / misunderstanding possess the identifying determination note identification means and the determination noted as being unknown,
The problem presenting means includes a problem included in the learning unit having the largest change amount of the understanding degree probability before and after recalculating the understanding degree probabilities of all other learning units not yet asked by the understanding degree probability updating means. A device characterized by the following questions . A device testing method for performing user-adaptive testing,
The apparatus has a learning unit storage unit that stores in advance a directed graph in which learning units including questions and answers are connected as nodes, and the learning units are connected with arcs of prior understanding probabilities,
The device is
For each user, a first step of calculating a subsequent understanding probability in all other learning units that have not been answered for correct / incorrect questions included in the existing learning unit;
For each user, a second step of presenting questions included in an unquestioned learning unit;
A third step of determining, for each user, correct / incorrect answers to the question;
When a correct answer is given to a question of a learning unit whose ex-post understanding probability is lower than the determination threshold, or when a wrong answer is given to a question of a learning unit whose understanding probability is higher than the determination threshold, the learning unit A fourth step for identifying the judgment / notice as unclear / misunderstood
Run
In the second step, before and after recalculating the understanding probabilities of all other learning units that have not been asked in the first step, the problems included in the learning unit having the largest change in the understanding probability are To answer questions
A method for testing an apparatus, characterized in that

Images