JP2021162732A - Subject recommendation system - Google Patents

Abstract

To solve the problem that a conventional learning support system cannot create questions with high accuracy because a plurality of elements are not considered, such as different field elements as to whether to previously know background knowledge of questions, formative elements of question creation formats and the number of blanks, and the length of question sentences and the difficulty of words in that questions to be solved next are recommended depending on "only" grammatical items.SOLUTION: A subject recommendation system is provided, which extracts questions to which the same tag as a tag representing a theme which a user requests for recommendation from a question database by using the question database for managing a plurality of questions, tags, answers of the questions and answer contents, and an answer state database for managing information on correctness and wrongness of answers to the questions of a plurality of respective users, and makes questions having wrong answers made by different users similar to the user among the questions to be questions to be recommended to the user.SELECTED DRAWING: Figure 1

Description

The present invention relates to a system that recommends a task for learning, and more specifically, for a self-study system, not only a task that a learner is explicitly not good at but also a task that is considered to be potentially weak. It relates to a problem recommendation system that brings about a high learning effect by making recommendations based on learning results using information recommendation technology using knowledge and artificial intelligence technology.

Various self-study support systems have been developed since the dawn of computer development, and it has been a long time since services that provide appropriate tasks according to the learner's level have been put into practical use. Since personal computers, smartphones, tablets, etc. can be used to access these services, various services are provided for the purpose of preparing for various exams that require a lot of time, and high school entrance exams are taken. Typical examples are preparations for university entrance exams, preparations for various qualification exams, and preparations for language exams.

Since it is necessary to obtain a certain number of points in each case, it is necessary to acquire the knowledge and academic ability to answer the correct answers to the questions given in these exams. Therefore, the purpose of the above service is to support the efficient acquisition of knowledge and academic ability.

It is not always possible to answer the correct question next time, but it is more likely to answer the wrong question next time. Therefore, it is important to have a learning method that changes the error to the correct answer next time by repeatedly tackling only the wrong problem in order to effectively utilize the limited learning time.

As a result, many services have a re-learning feature for false problems. A certain effect can be obtained by re-questioning the wrong question as it is, but if you re-learn immediately after answering, there is a possibility that you will choose the correct answer even though you do not understand the essence due to the relationship of choices etc. Cannot be excluded. Therefore, it is effective to encourage the user to learn the theme by re-questioning another question that belongs to the same theme as the theme that the wrong question was really asking.

Japanese Patent No. 6618226 JP-A-2017-134184

I would like to consider this point by taking English learning as an example. In Japan, when learning English, 1 basic grammar action verbs and stative verbs, present tense and present progressive tense, past tense and past progressive tense, etc.
2 Auxiliary verb recommendation may, can and must, auxiliary verb and perfect form, idiomatic expression of auxiliary verb, etc.
3 Subjunctive mood Subjunctive mood Subjunctive mood, subjunctive mood, subjunctive mood, etc. 4 Subjunctive mood It is common to study systems such as gerund usage and negative forms of gerunds in order.
Therefore, the examination related to English also has a strong viewpoint of confirming whether or not the grammar has been mastered. Many English-related question books and learning support systems also consist of multiple questions that are categorized from the perspective of which grammar question is being asked. For example, the following problems.

Question 1 Select the option that is considered appropriate for () below.
The wedding party () a lot of money. There were a lot of guests, and the hotel was the best in the city.
1 must have cost
2 cannot have cost
3 could not cost
4 will cost

In the case of this problem, the grammatical item "auxiliary verb + perfect form" related to must have cost, which is the correct answer option that can be entered in the blank, is the theme that indicates the problem, and this problem belongs to 2 auxiliary verbs in the above system. It means "auxiliary verb and perfect form".

However, the point of answering this question lies in the second sentence "were" and "was", and it is essential to understand that these are the past tense. That is, a user who does not understand the past tense cannot reach the correct answer to this problem even if he / she understands the auxiliary verb.

When the conventional learning support system re-questions, this point is ignored, and the next problem to be solved is recommended by relying "only" on the auxiliary verbs and perfect forms, which are grammatical items related to the answer column. Therefore, we only recommend problems that are not essential countermeasures. Therefore, it remained impossible to improve academic ability by answering questions about the past tense, and it was not efficient from the viewpoint of improving academic ability and acquiring knowledge.

Furthermore, when the teacher judges the achievement status of the student, not only the grammatical theme of the question, but also other field factors such as whether or not the background knowledge of the question is known in advance, the question format and the number of blanks are formal. Factors, the length of the question sentence, the difficulty of the word, and the psychological factors such as whether the answer was confused or unlucky with a pencil roll are also taken into consideration. It was not considered.

By the way, in recent years, artificial intelligence has been put into practical use on a large scale. Pattern matching is a specialty of artificial intelligence, but in learning support systems, it is used to calculate which pattern the user's wrong answer belongs to, that is, a problem similar to the wrong answer. It is a promising technology because it is difficult to understand at first glance when using artificial intelligence, but it has an important relationship and has the potential to recommend problems that contribute to understanding the theme. On the other hand, the weakest point of information recommendation technology using artificial intelligence technology is that it is not possible to specify the "reason" why the problems were calculated to be similar, and the "knowledge" of the expert is the reason. The point is that it is difficult to reflect it well in the output result.

Therefore, in the present invention, a problem recommendation providing system, that is, a problem recommendation providing system, characterized in that a tag is attached to each problem, that is,
A question database that manages multiple questions, tags, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a problem recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A problem recommendation system characterized by being a problem.
I will provide a.

The problem recommendation system of the present invention makes it possible to classify each problem from various viewpoints by attaching a plurality of tags to each problem, which is really necessary for the user to acquire knowledge, not from a uniform viewpoint. It is possible to efficiently provide problems belonging to various themes. Furthermore, by recommending problems that similar users who are judged by information recommendation technology using artificial intelligence technology are not good at, problems that users are likely to be bad at and problems that users have not yet answered. Only can be recommended, and the efficiency of knowledge acquisition can be greatly improved. In particular, in language learning, which has often depended only on the system of grammar, it is possible to extremely improve learning efficiency by making recommendations from a viewpoint other than grammar.

Outline of the system for carrying out the present invention Flow chart explaining the operation of the present invention

The subject recommendation system 1 of the present invention will be described with reference to the following figures. Problem The recommendation system 1 is composed of a distribution server 2 and a client terminal 3. FIG. 1 is a diagram showing an outline of the task recommendation system 1, and FIG. 2 is a flowchart diagram.

Here, in the task recommendation system 1 of the present embodiment, long sentences, short sentences, multi-limb selection formulas, alternative formulas, and free descriptions are used for various grammars possessed by English, with knowledge in various fields including language as a learning target. It functions as a system for solving problems that are asked in various question formats such as formulas, and recommends problems that enable efficient language learning. In the following, the explanation will focus on language learning, especially English.

As shown in FIG. 1, the task recommendation system 1 of the present embodiment is connected to the Internet 1000, can store the problem collection contents 4 useful for learning English according to the difficulty level, and can distribute the problem collection contents 4 through the Internet 1000. It is composed of a distribution server 2 and a client terminal 3 capable of displaying the problem collection content 4 distributed by the distribution server 2 to the user 1001 and inputting the answer of the user 1001. The distribution server 2 is installed by a company that implements an English learning support service that operates the task recommendation system 1, and the client terminal 3 is managed by the user 1001, but the client terminal 3 is also used by the company in a classroom or the like. It may be installed and managed collectively. An intranet may be used as an alternative to the Internet, or a stand-alone system that does not have a network connection may be used.

The distribution server 2 is hardware composed of CPU, RAM, storage such as HDD and SDD, network card, power supply, etc., and various English-related information can be obtained by cooperating with software executed on the hardware. The problem database 21 in which the problem collection content 4 is managed, the web page generation means 22 that dynamically generates the problem 41 extracted from the problem database 21 as a web page, and the communication means 23 that transmits / receives to / from the client terminal 3. , The answer status database 24 that manages the answer status from the user 1001, and the recommendation system 25 that extracts questions similar to the question 41 that the user erroneously answered. Further, a pointing device such as a keyboard, a mouse, and a track pad for enabling control of the distribution server 2 itself, a display for displaying information of the distribution server 2 itself, and the like may be provided. The existing DBMS may be used as the database, or alternative software having similar functions may be used. Although the Web page generation means 22 generates HTML, other languages such as PDF and PS may be used, and the problem 41 may be processed into information that can be recognized by the user.

When the distribution server 2 receives the problem collection distribution request from the client terminal 3 operated by the user 1001 via the communication means 23, the distribution server 2 extracts the problem 41 according to the theme for which the user 1001 requests distribution from the problem database 21. The extracted problem 41 is processed into Web page display data that can be rendered as a problem display screen by the Web page generation means 22, and then distributed to the client terminal 3 via the communication means 23.

The client terminal 3 is a personal computer, a tablet-type personal computer, a smartphone, etc., and more specifically, hardware composed of a CPU, RAM, storage such as HDD and SDD, a network card, a power supply, a display, a speaker, and the like. Yes, by collaborating with the software executed on the hardware, the problem 41 related to English is displayed in addition to the display means 31, the problem collection to the distribution server 2, the distribution request, the recommendation request, and the like. It functions as an input means 32 that accepts various inputs such as accepting answers to the question 41, a communication means 33 that can communicate with the distribution server 2, and a storage means 34 that temporarily stores the answer status of the user 1001. The display means 32 is configured by the cooperation of the display, the speaker, and the software, and presents the content of the problem 41 to the user 1001 by visually displaying the problem 41 or reproducing the problem 41 audibly.
The input means 32 may be a pointing device such as a keyboard, a mouse, or a track pad, or may be a touch panel integrated with the display means. In addition to these, the input means 32 also has a timekeeping function for measuring the time the user has in answering, in addition to a camera that catches the line of sight of the user 1001 and a sound collecting microphone that catches the user's voice.

The most different point of the problem recommendation system 1 from the conventional technique is that each problem managed by the problem database 21 is tagged by an educational expert. The tag is an extraction of the elements necessary for the answerer to judge the correct answer and the elements that influence the answering activity (hereinafter referred to as "answer elements") when answering the multiple-choice question. , Multiple tags are attached to one problem.

Tags are managed according to difficulty level as a tag book composed of three attributes, a grammatical attribute, a formal attribute, and a psychological influence element. Grammar attributes are set according to the major classifications of English grammar such as sentence patterns, voices, auxiliary verbs, infinitives, gerunds, participles, relatives, comparisons, assumptions, conjunctions, and grammars. Middle classification and minor classification are set for each of these classifications. For example, middle classification "must" and middle classification "can" are arranged under the major classification "auxiliary verb", and the middle classification "must" is also The sub-category "tense" and the sub-category "possibility" are arranged. In addition, the major classification "tense" includes the middle classification "present tense", the middle classification "past tense", the middle classification "great past", and the middle classification "past tense" includes the minor classification "basic". Sub-category "Other" etc. are arranged. Each subclass is given a degree of importance in view of the importance of its grammatical items. For example, the major classification "auxiliary verb", the middle classification "must", and the minor classification "possibility" are given a degree of importance of 3. As a tag, it is expressed in the form of "auxiliary verb-must-possibility, importance 3".

As a primary grammar, each question is tagged with the elements that are considered to be directly necessary to fill in the answer column, in descending order of relevance. Furthermore, as the secondary grammar, tags that are considered necessary for the answer other than the primary grammar are attached in order from the one with the highest degree of relevance.

The gentlest difficulty 1 tagbook does not include high tags related to advanced grammar such as subjunctive mood and voice. This is because these grammatical elements are not asked in the problem assumed by difficulty level 1. However, the tagbooks with moderate difficulty 2 and the most difficult difficulty 3 also include tags belonging to the subjunctive mood and voice. Even with the same tag, the importance changes according to the difficulty level. For example, "tense-past-be verb" is the most important importance 3 in the difficulty 1 tagbook, but the least important importance 1 in the difficulty 3 tagbook. This is because it is considered that the past tense of the be verb has been definitely mastered for the user who works on difficulty level 3.

Formal attributes are set according to the major classification that focuses on the formal aspect rather than the content of the question, such as the question format, question format, answering elements, presence / absence of Japanese, number of blanks, number of words in the question sentence, and sentence format. Minor classifications are set for each of these classifications. For example, the major classification "question format" includes the minor classification "symbol type", the minor classification "descriptive type", the minor classification "sorting type", and "others". , Etc. are arranged, and the question format is "sentence selection type", "vacant space filling type", etc. Classification "word supplement", minor classification "word correction", minor classification "sort (whole sentence)", minor classification "sort (part of sentence)", etc. are arranged, and the major classification "number of blanks" is small Classification "0", minor classification "1", minor classification "2 or more", major classification "number of words in problem sentences" is minor classification "less than 20 words", minor classification "20 words or more and less than 40 words", minor classification " "40 words or more", the major classification "sentence format" is the minor classification "normal sentence", the minor classification "dialogue sentence" and the like. As a tag, it is expressed in a format such as "element to answer-word supplement", "number of words in question sentence-41 words or more". As for the tags belonging to the format attribute, all the tags corresponding to each problem are attached.

Psychological influence elements are composed of answer grounds, question sentence genres, word difficulty levels, special question formats, and the like. The major classification "answer basis" includes the minor classification "grammar", the minor classification "word usage", the minor classification "vocabulary", the minor classification "context", "English logic", etc., and the major classification "genre of question sentences". The minor classification "abstract sentences", the minor classification "social science", the minor classification "foreign culture", the minor classification "business", etc. are arranged in the major classification "word difficulty", and the minor classification "1st grade". "Above", minor classification "quasi 1st grade", minor classification "2nd grade", minor classification "less than 2nd grade", etc. are arranged. They are "proper nomenclature", sub-classification "No error type", and sub-classification "blank at the beginning of sentence". In addition, the major classification "special question format" has a different perspective from the question format as a formal attribute, and it is highly difficult to obtain a psychologically correct answer recognized from experience and research associated with many years of English conversation instruction. It is about the question format to be asked. In other words, from the viewpoint that accurate knowledge is required due to the question format, it is difficult to associate, etc., so the recognition that the difficulty is high comes first, and the weak consciousness is stimulated and it is easy to fall into a negative psychological state. Is tagged from. As a tag, it is expressed in a format such as "answer basis-word usage". As for the tags belonging to the psychological influence elements, the tags that may cause the user's weakness and the like are attached in order from the one with the highest degree of relevance.

Hereinafter, how to attach tags will be further explained through several examples.
Question 1 Select the option that is considered appropriate for () below.
The wedding party () a lot of money. There were a lot of guests, and the hotel was the best in the city.
1 must have cost
2 cannot have cost
3 could not cost
4 will cost

In question 1, the primary grammar is tagged with "auxiliary verb-must-tense, importance 3" and "auxiliary verb-must-possibility, importance 3", and the secondary grammar is "tense-past tense-basic, important". Degree 2 "is added. Formal attributes are "question-vacant space filling type", "answer-correct phrase", "answer number-1", "question word number-21 words or more and less than 40 words", "with or without Japanese-nothing", "sentence" Format-normal sentence "is added. As a psychological influence factor, "answer basis-context" is attached.

Question 2 When rearranging words so that they have the same meaning as Japanese sentences, select the one word that is missing.
"It's almost impossible to get your homework done by tomorrow."
It is (impossible, finish, to next) the homework tomorrow.
1 the
2 with
3 of
4 to

In question 2, the tag "infinitive-noun usage-special expression, importance 2" is attached as the primary grammar, and there is no secondary grammar. As formal attributes, "question-sort type", "answer-correct phrase", "answer number-1", "question word number-20 less than", and "presence / absence of Japanese-existence" are added. As a psychological influence factor, "answer basis-vocabulary" is attached.

Problem 3 Muhammad Yunus, an idealist who wanted to ensure that loans are available to the poor, set up Grameen Bank; most of its lending is () such as street vending and farming.
1 for income-generating activities
2 making significant profits
3 to help excluded but motivated people
4 an alternative to large bank funding

In question 3, the primary grammar is tagged with "sentence pattern-2 sentence patterns-type of complement, importance 1", and the secondary grammar is tagged with "others-vocabulary-idiom". Formal attributes include "question-fill in vacant spaces", "answer-correct phrase", "number of answers-1", "number of question words-20 words or more and less than 40 words", "form-normal sentence", "Japanese-" No "is added. As psychological influence elements," answer basis-context "," answer basis-English logic "," sentence content-foreign culture ", and" special question format-proprietary nomenclature "are added.

This problem cannot be answered without knowledge of two sentence patterns from the viewpoint of English grammar knowledge, and the viewpoint of grammar has been emphasized in conventional English learning. If added, grammatical knowledge such as "apposition", "subordinating", and "insertion" is also required at the same time, so that viewpoint has also been taken into consideration.

However, although it is considered that the correct answer can be surely given if the English grammar knowledge exists at a sufficient level, there may be cases where the correct answer can be given even if the person does not necessarily have sufficient grammar knowledge. In question 3, if you have knowledge of microfinance implemented by Grameen Bank, it is easy to infer the meaning of words that are a little unfamiliar to Japanese learners of English, such as street vending and farming at the end, and the whole sentence. The image of is also easy to understand. The bank's governor, Muhammad Yunus, has won the Nobel Peace Prize and has been widely reported in Japan, so it is highly possible.

In this way, if you have background knowledge, you can correctly estimate the content of the question sentence if you have some English ability, especially vocabulary ability, and as a result, it often leads to the correct answer. Found by them. Learners who achieve a high correct answer rate by using such background knowledge while the present inventors are teaching English, and by analogizing and estimating the content of the entire sentence using the English knowledge of the partial part as a foothold. The existence of this type has also been confirmed.

In the present invention, focusing on this point, it is decided to tag items that have no relation to grammatical knowledge and English knowledge represented by vocabulary, but have strong elements that lead to correct answers. As a result, by advancing the analysis of the tag, the strengths and weaknesses of this type were determined, and more efficient recommendations were made possible. In other words, even if the answer is correct, it is possible to recommend the same type of grammar problem if it is considered that there is not enough grammar knowledge.

Question 4 Of the following questions (a) to (d), if there is a part that contains an incorrect English expression, select one of (a) to (d), and if there is no error, ( e) Select No error.
(A) Having been in disagreement (b) for some time , the two sides made a temporary pact, but as everyone knows (c) all too well , the quarrel is (d) far from over. (E) No error

In question 4, the tag "participle-participle syntax-tense, importance 2" is attached as the primary grammar, and there is no secondary grammar. Formal attributes include "question-underlined", "answer-wrong word", "answer number-1", "question word number-21 words or more and less than 40 words", and "with or without Japanese-nothing". NS. Psychological influence factors include "answer basis-context", "answer basis-English logic", "sentence content-foreign culture", "special form-proper noun", and "special question form-No error type". NS.

Here, explaining the No error type, it has been a long time for applicants that there is a type of learner whose ability to grasp the content and answer to a question is lowered and the correct answer rate is clearly lowered because No error is included in the options. It is known from the experience and research of teaching English learning. In addition, by including No error, more accurate knowledge is required and the difficulty level increases, so the correct answer rate does not decrease, but the recognition that the difficulty level is high comes first, and the negative psychology that accompanies the weakness consciousness. It is analyzed that the action comes first and the correct answer rate decreases. Similar to No error, the same applies to elements such as long problem sentences, many choices, and long choices, and the feature of the present invention is that these are tagged as psychological influence factors.

It is extremely beneficial for these types of learners to have the experience of eliminating negative non-English psychological effects, that is, the experience of repeatedly and intensively solving low-English problems with the same question format. .. Therefore, in the present invention, a tag of a psychological influence element is introduced in order to efficiently recommend a problem having a common psychological influence element. With the introduction of the psychological influence element tag, it is possible to create a problem set in which the psychological influence element is common but the learning contents related to English such as grammatical knowledge and word knowledge are different, so it is easy to lower the English level.

Question 5 Select the option that you think is appropriate for () below.
After leaving the room, he suddenly remembered that he () the important papers behind.
1 leaves
2 has left
3 left
4 had left

This problem is tagged with "tense-past form-matching tense" and "tense-past form-perfect tense", but in the difficulty level 1 tag book, "tense-past form-matching tense" , Importance 4 ”,“ Tense-Past Form-Perfect Form, Importance 4 ”, Difficulty 3 tagbook,“ Tense-Past Form-Tense Match, Importance 1 ”,“ Tense-Past Form-Completed The importance is set to differ depending on the difficulty level, such as "shape, importance level 2".

By changing the importance according to the change in the difficulty level, it becomes easy to recommend the optimum problem according to the learning situation of the user. In order to answer question 5 correctly, it is essential that the meanings and usages of "sequence of tenses" and "past perfect" are completely mastered. However, it is difficult for beginners to intuitively understand these English grammar items because such grammar operation does not exist as a serious problem in Japanese grammar. To understand it, it is necessary to understand English grammar theory and to have some experience (proficiency) in English.

However, since these items are frequent items for advanced users, they can be handled semi-unconsciously without stopping. That is, it is of very low importance to advanced players. Therefore, there is little need to solve this problem, and it is rational to lower the priority of recommendations, and this is achieved by setting a tag book according to the difficulty level.

In the problem database 21, the above-mentioned tags are attached to each of the problems 41 and managed. That is, a question ID, a question sentence, an answer to the question, a tag related to one or more primary grammars, a tag related to one or more secondary grammars, a tag related to formal attributes, and a tag related to psychological influence elements.

In the answer status database 24, a question ID for each user and whether the question is answered correctly, incorrectly, or unanswered are stored. The specific data format can be freely designed using known technology. For example, regarding the answer status of a certain user, if the component is 0, the answer is unanswered, 1 is the correct answer, and -1 is the wrong answer. Good).

Next, the processing flow of each of the distribution server 2 and the client terminal 3 in the problem recommendation system 1 will be described with reference to FIG. Here, it is assumed that a certain user 1002 is the answerer.

The distribution server 2 manages the problems tagged in advance in the problem database 21 so that they can be transmitted to the client terminal 3 through the Web page generation means 22 and the communication means 23. At this time, the problem database 21 is provided with an ID unique to the client terminal 3, an authentication system using a user ID and a password, and the like. However, this does not apply if security can be ensured, such as in a closed environment that does not connect to the general Internet.

The user 1002 logs in to the client terminal 3, selects a difficulty level through the input means 32, and makes a problem distribution request for a problem belonging to the difficulty level. The client terminal 3 transmits the problem distribution request to the distribution server 2 through the communication means 33 together with the difficulty level. At this time, the tag may be selected in addition to the difficulty level. For example, you may select a question related to the tag "tense" of difficulty level 2.

When the problem distribution request is received from the client terminal 3, the distribution server 2 selects the problem according to the problem distribution request, reads the selected problem from the problem database 21, and displays it on the client terminal 3 by the Web page generation means 22. After making it suitable data, it is transmitted to the client terminal 3 through the communication means 33. The question distribution request is managed by the question database, such as questions belonging to a certain difficulty level, those limited to grammar items, those limited to multiple grammar items, and limited to a certain question format. It can be freely selected according to the type of tag.

The client terminal 3 displays a problem on the display means 31 using the data from the distribution server 2 received via the communication means 33. The answer input as a result of the user 1002 examining the problem is received through the input means 32, and the answer is temporarily stored in the storage means 34.

When the user 1002 inputs all the answers, the display means 31 starts the display for receiving the instruction to start scoring. When the user 1002 inputs an instruction to start scoring through the input means 32, the client terminal 3 scores the answer input by the user 1002 and displays the scoring result on the display means 31. At the same time, the scoring result is transmitted to the distribution server 2 through the communication means 33. The distribution server receives the scoring result through the communication means 23, and records the question ID of each question and the correctness in the answer status database 24 together with the user ID of the user 1002.

At this time, the answer stored in the storage means 34 is transmitted to the distribution server 2 through the communication means 33, and the answer received through the communication means 23 is scored in the distribution server 2 to obtain a scoring result and stored in the answer status database 24. The recording may be transmitted to the client terminal 2 through the communication means 23, and the scoring result obtained through the communication means 33 may be displayed on the display means 31. In that case, it is not necessary for the client terminal 3 to hold the data required for scoring.

Along with the display of the scoring result, the display means 31 starts displaying for accepting the recommendation request. When the input means 32 receives the input of the recommendation request by the user 1002, the client terminal 3 transmits to that effect to the distribution server 2 through the communication means 33. When the distribution server 2 receives the recommendation request through the communication means 23, the tag attached to the problem ID that was incorrect most recently is specified as the recommendation tag candidate 43, and the recommendation tag candidate 43 is specified as the recommendation tag candidate 43 through the communication means 33 as the client terminal. It is transmitted to 3 through the communication means 23.

The client terminal 3 displays the recommendation tag candidate 43 received through the communication means 33 on the display means 31. The user 1002 comes into contact with the recommendation tag candidate 43 displayed on the display means 31, and uses the recommendation tag 44 from the recommendation tag candidates 43 based on what kind of problem the recommendation is required. To select. The client terminal 3 transmits the recommendation tag 44 to the distribution server 2 through the communication means 33. Alternatively, the user 1002 may freely select the recommendation tag 44, or the recommendation tag candidate 43 may be selected as the recommendation tag 44 as it is.

At this time, when the input of the recommendation request is received by the input means 32 in the client terminal 3, the tag attached to the most recently incorrect problem ID is used as the recommendation tag candidate 43 without communicating with the distribution server 2. It may be displayed on the display means 31 and selected by using the input means 32. In this case, it is necessary to distribute the information about the problem ID and the tag to the problem distributed to the client terminal 3 in advance.

When the distribution server 2 receives the recommendation tag 44 through the communication means 23, the problem database 21 sets the problem to which the recommendation tag 44 is attached and the user 1002 has not yet answered as the recommendation problem candidate 44. Extract.

The distribution server 2 calculates a set of users 1003 whose erroneous answer status is considered to be similar to that of the user 1002. This calculation may be performed by applying an information recommendation technique using a widely known artificial intelligence technique or the like. That is, the network composed of a plurality of layers is trained by using the data related to the user's wrong answer as an input signal. Then, using the learned network, it is sufficient to calculate the users who are considered to be similar by inputting the data regarding the wrong answer of the user 1002. The similarity between users may be expressed by setting a vector for each user with the situation regarding the wrong answer of each user as an element and expressing it by the angle formed by the vector.

Since a huge amount of calculation is required for these calculations, it is necessary for the software to automatically perform the calculation using hardware resources such as CPU and memory in the distribution server. Since a large amount of computing power is required even if such hardware resources are used, it is desirable to perform the calculation during a time period when the processing capacity of the distribution server 2 is sufficient. In that case, the calculation result is stored in the distribution server 2 as a similarity table, and when it becomes necessary to select a user similar to a certain user, the similarity is read from the similarity table instead of being calculated each time. To do so.

This calculation requires each user to solve a certain number of common problems. Otherwise, it is difficult to calculate similarities and dissimilarities between users. Therefore, each user may be encouraged to solve an essential problem composed of a plurality of fields and a plurality of difficulty levels, and may be used as a basic material for calculating the similarity between users. .. Even if there are no essential questions, the similarity can be calculated as long as a sufficient number of common questions are answered by each user, so it may be left to the user to the last.

Next, for each question of the recommendation question candidate 44, if each user belonging to the set of users 1003 gives an incorrect answer, a value multiplied by the degree of similarity with the user 1002 is added, and no incorrect answer is given. In this case, no points are added, and the quotient divided by the user set 1003 after performing the calculation for all the users belonging to the user set 1003 is defined as the incorrect answer rate. That is, the weighted average of the incorrect answer status of each question is obtained by weighting the similarity of the users. If a user answers the same question multiple times, the above points are added each time.

A certain number of questions from the one with the highest error rate are determined as the recommendation question 42. At this time, the importance may be added in addition to the wrong answer rate. For example, for a question of high importance, the importance can be added to or multiplied by the wrong answer rate, and if the user has specified a tag, only the importance given to the tag is incorrect. You can add or multiply by the answer rate. As a fixed number, various numbers can be set, such as a number suitable for review, a number set by the user according to his / her own taste, and a number designated by an expert as suitable from the learning situation.

The server 2 reads the problem from the problem database 21 using the problem ID determined in the recommendation problem 42, makes the data suitable for display on the client terminal 3 by the Web page generation means 22, and then passes the client terminal through the communication means 33. Send to 3.

At that time, problems from completely random fields may be mixed. This is not only effective from the viewpoint of preventing so-called overfitting, but also significant from the meaning of providing new awareness material regarding English learning by recommending unexpected problems by the user. Problems that are not random but have some relevance, for example, problems that users belonging to the user set 1003 often answer incorrectly and do not have the recommendation tag 44 attached, or problems that belong to the user set 1003 It is extremely useful to mix questions that are often mistakenly answered and that do not have the recommendation tag candidate 43 attached.

As described above, the essence of the present invention is that the distribution server 2 can specify the field in which the user requests a recommendation, and the problem belongs to the field and has not yet been answered by the user. Among the problems, it is possible to recommend a problem that is often mistaken by another user who is similar to the user. Although the calculation is an amount of calculation that cannot be processed by humans, it can be performed by the cooperation of a computer that is widely used at present, that is, a hardware resource and software for operating the hardware.

Although the recommendation calculation itself requires hardware resources and software as described above, the rest can be modified without losing the essence of the present invention. For example, it is conceivable that the user or a person who asks a question to the user directly operates the distribution server 2, obtains information on a tag or difficulty level for which a recommendation is requested, and makes a recommendation based on the information. .. In this case, since a computer is indispensable for the calculation using the linear network, it is indispensable to use the hardware resources for the distribution server 2, but paper may be adopted as the client terminal 3. ..

Although the embodiments of the present invention have been described above, some or all of the embodiments of the present invention will be described as described in the following appendices.
[Appendix 1]
A question database that manages multiple questions, tags, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a problem recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A problem recommendation system characterized by being a problem.
[Appendix 2]
Computer,
A question database that manages multiple questions, tags, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a program that functions as a task recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A program characterized by performing the processing in question.
[Appendix 3]
A question database that manages multiple tags, questions, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a data structure used in the problem recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A data structure containing tags, questions, correct answers to the questions, and answer contents used in the processing of the question.
[Appendix 4]
The problem recommendation system described in Appendix 1,
A task recommendation system characterized in that a plurality of tags are attached to each of the above problems.
[Appendix 5]
The problem recommendation system described in Appendix 4,
A task recommendation system characterized in that the plurality of tags include psychological influence element tags attached from the viewpoint of the answer basis, the genre of the question sentence, the difficulty level of the word, or a special question format.

1 Problem recommendation system 2 Distribution server 21 Problem database 22 Web page generation means 23 Communication means 24 Answer status database 25 Recommendation system 3 Client terminal (CPU,
31 Display means 32 Input means 33 Communication means 34 Storage means 4 Question collection content 41 Delivered question 42 Recommendation question 43 Recommendation tag candidate 44 Recommendation tag 45 Recommendation question candidate 1000 Internet 1001 User 1002 Answerer User 1003 Wrong answer A set of users with similar situations

Claims (5)

A question database that manages multiple questions, tags, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a problem recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A problem recommendation system characterized by being a problem. Computer,
A question database that manages multiple questions, tags, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a program that functions as a task recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A program characterized by performing the processing in question. A question database that manages multiple tags, questions, correct answers to the question, and answer contents,
An answer status database that manages information about the correctness of the answer to the above question for each of multiple users,
By learning in advance information recommendation technology using artificial intelligence technology or the like using information on the correctness of the answers of some of the plurality of users as an input signal, information on the correctness of the answers of a certain user can be obtained. An arithmetic device characterized in that when it is input to an information recommendation technique using the artificial intelligence technology or the like, it is possible to output a user who has information on the correctness of an answer similar to that of the certain user.
It is a data structure used in the problem recommendation system equipped with
Using the arithmetic unit in advance, the similarity is calculated between the certain user and the plurality of users to create a similarity table.
A problem with the same tag as the tag expressing the theme for which the user wants a recommendation is extracted from the problem database, and the problem that the user has not yet answered is extracted from the problem database.
Among the extracted problems, a plurality of other users similar to the certain user are selected from the similarity table, a plurality of problems are selected from the problems that the other user is erroneous, and a recommendation is made to the certain user. A data structure containing tags, questions, correct answers to the questions, and answer contents used in the processing of the question. The problem recommendation system according to claim 1.
A task recommendation system characterized in that a plurality of tags are attached to each of the above problems. The problem recommendation system according to claim 4.
A task recommendation system characterized in that the plurality of tags include psychological influence element tags attached from the viewpoint of the answer basis, the genre of the question sentence, the difficulty level of the word, or a special question format.

Images