KR20230063759A - Selecting questions method for english proficiency test - Google Patents

Abstract

The present invention relates to a method of selecting questions for an English proficiency test. The method of selecting the questions in a system for the English proficiency test comprises the steps of: obtaining source data by using a data server in which the previously performed English questions are stored; constructing, by the data server, a dataset by analyzing the source data into a plurality of data information including keywords, themes, difficulty levels, question format, fingerprint structures, and fingerprint format; grouping and classifying, by the data server, the questions with similar problem types in the constructed dataset; grouping and classifying, by the data server, the questions with similar difficulty levels in the constructed dataset; and randomly selecting, by the data server, the questions in consideration of the classified similarity and difficulty level. According to at least one of the embodiments of the present invention, it is possible to analyze the similarity and predict the difficulty level of a question set extracted from a question bank.

Description

How to select questions for the English proficiency test {SELECTING QUESTIONS METHOD FOR ENGLISH PROFICIENCY TEST}

The present invention relates to a method for selecting questions for an English proficiency test, and more specifically, to a method for grouping items according to similarity and difficulty of items and selecting items based thereon.

The biggest problem with domestic English education failing to overcome the high-cost and low-efficiency structure is that education that does not fit the learner's level is being conducted without an objective and accurate diagnosis.

Currently, questions in the English proficiency test are biased according to the inclination and intention of the examiner, or cause an imbalance in topics and difficulty. Therefore, there is a disadvantage that is not efficient in presenting a learning plan according to the grade result.

The present invention aims to solve the foregoing and other problems. Another purpose is to provide a method for selecting questions in the system for the English proficiency test.

According to one aspect of the present invention to achieve the above or other objects, in a method for selecting questions in a system for an English proficiency test, obtaining source data by using a data server in which previously performed English questions are stored; constructing a dataset by the data server by analyzing the source data into a plurality of pieces of data information including keywords, themes, difficulty levels, question formats, fingerprint structures, and fingerprint formats; Grouping and classifying, by the data server, questions having similar problem types in the constructed dataset; Grouping and classifying, by the data server, questions having similar difficulties in the constructed dataset; The data server randomly selects questions in consideration of the classified similarity and level of difficulty, and the English questions include prompts that provide situational understanding or background knowledge to solve the problem, and content to be asked. Includes a question defining the nature of the item, and a extractor including a correct answer, classifies the passage according to a plurality of themes, and classifies the question according to a plurality of question types. and classifies the reason why the non-correct answer among the options is not correct into a plurality of types, and assigns a predetermined score to an item belonging to the same group for each of the reasons for the passage, question, and non-option, and then the fingerprint, A question selection method for an English proficiency test may be provided, in which questions and options are judged as similar questions when a predetermined score or more is obtained.

According to one aspect of the present invention, the classification criteria for the fingerprint are Actions, Activities, Animals, Art, Careers, Clothing, Colors, It may include at least one of Food, Emotion, Entertainment, Health, Music, and Nature.

According to one aspect of the present invention, the questions are classified according to a plurality of question types, but the classification criteria for the question types include finding answers appropriate for the Who question, finding answers appropriate for the What or Which question, finding appropriate answers for the Where question, It may include one or more of searching for an answer appropriate for a When question and finding an appropriate answer for a How question.

According to one aspect of the present invention, the classification criteria for the reason why the non-correct answer is not correct among the options are ambiguous pronouns, confusing words, keyword related expressions, and keyword related At least one of keyword unrelated, missing keywords, identical expression (Overlap), wrong description (Wrong description), wrong meaning (Wrong meaning), wrong sequence (Wrong sequence), wrong situation (Wrong situation) may contain one or more.

According to one aspect of the present invention, when the fingerprint is provided in the form of a picture, the picture may be replaced with a description of information and pictures necessary for problem solving in sentences or text.

According to an aspect of the present invention, after the step of randomly selecting an item by the data server, the step of transmitting the item selected by the data server to the test taker terminal may be further included.

The effect of the problem selection method for the English proficiency test according to the present invention is described as follows.

According to at least one of the embodiments of the present invention, similarity analysis and difficulty prediction of item sets extracted from a question bank are possible, and different items can be provided according to test takers in real time.

According to at least one of the embodiments of the present invention, similarity analysis and difficulty prediction AI are introduced using an AI platform, and non-face-to-face by generating a set of items (SET) that can ensure quality enough to replace regular exams It has the advantage of preoccupying the test market.

According to at least one of the embodiments of the present invention, it is possible to analyze the type, difficulty, and correct answer rate of the questions presented, thereby efficiently presenting an English learning direction suitable for each test taker, and realizing personalized English learning. There are advantages to

According to at least one of the embodiments of the present invention, when an English proficiency question is set, the question set by the setter is compared with the previously set question, and AI presents an analysis of difficulty, type, and correct answer rate, thereby systematizing and beneficial to efficiency.

A further scope of the applicability of the present invention will become apparent from the detailed description that follows. However, since various changes and modifications within the spirit and scope of the present invention can be clearly understood by those skilled in the art, it should be understood that the detailed description and specific examples such as preferred embodiments of the present invention are given by way of example only.

1 is a flowchart of an item selection method for an English proficiency test related to the present invention.
2 is an example of a question related to the present invention.
3 is a system diagram for an English proficiency test related to the present invention.

Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the accompanying drawings, but the same reference numerals will be assigned to the same or similar components regardless of reference numerals, and overlapping descriptions thereof will be omitted. In addition, in describing the embodiments disclosed in this specification, if it is determined that a detailed description of a related known technology may obscure the gist of the embodiment disclosed in this specification, the detailed description thereof will be omitted. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed in this specification, the technical idea disclosed in this specification is not limited by the accompanying drawings, and all changes included in the spirit and technical scope of the present invention , it should be understood to include equivalents or substitutes.

In order to prescribe an English study plan that is perfect for each individual, from the time the question is presented, the examiner must have a clear standard for what exactly he is trying to evaluate for each question. Therefore, if the examiner can analyze the similarity with past questions, difficulty, type, topic, subject intelligence, etc. with the help of AI when setting questions, then the level of completeness that can certainly provide English learning prescriptions to English learners A high level of service can be provided.

On the other hand, in order to more efficiently provide customized learning according to the test taker through test taker data analysis, it is necessary to upgrade the item setting system to provide a customized learning method to the learner based on the previously analyzed data when setting the item. there is.

To this end, it is necessary for an item setter to present existing item data according to the percentage of correct answers or according to the type of item when setting the item. The item data may be grade data evaluated at the same time with the same item for each level by freely selecting a level according to a cognitive level or an individual's English proficiency targeting infants, elementary school, middle school, high school, and college students nationwide in Korea.

In one embodiment of the present invention, after presenting the first question according to age or ability, the level of difficulty is adjusted according to whether or not there is a correct answer, a model capable of learning based on test result data is provided, and a pool of problems used is provided. ) is limited, the purpose is to provide similar questions for each question so that there is no problem with the difficulty level control algorithm while expanding the pool of questions indefinitely.

1 is a flowchart of a method for selecting English questions according to an embodiment of the present invention. Hereinafter, a question selection method for an English proficiency test such as a level test according to an embodiment of the present invention will be described with reference to FIG. 1 .

First, in the method of selecting questions in the system 200 for the English proficiency test, the step of obtaining original data using the data server 220 in which previously performed English questions are stored (S110), the data server 220 ) constructs a dataset by analyzing the source data into a plurality of data information including keywords, themes, difficulty levels, question formats, fingerprint structures, and fingerprint formats (S120), the data server 220 constructs the Step of grouping and classifying questions having similar problem types in the dataset (S130), step of the data server 220 grouping and classifying questions having similar difficulty in the constructed dataset (S140), the data server 220 ) includes a step (S150) of randomly selecting an item in consideration of the classified similarity and level of difficulty.

All of the steps S110 to S160 in one embodiment of the present invention are performed in the data server 220. The selection of the items may be performed by an algorithm such as AI, and may be randomly extracted in consideration of similarity and difficulty for each item. As shown in FIG. 2, the English questions include a prompt that provides situational understanding or background knowledge for solving the problem, a question that defines the nature of the question including the content to be asked, and a correct answer. It is made by including a distractor that includes. In some cases, the fingerprint is not a text but a picture or there is no fingerprint. In the case of a picture, the picture is replaced with the information that must be obtained to solve the problem and the description of the picture. That is, when there is only a picture without a fingerprint, the picture is described as a sentence or text and the degree of similarity is judged based on this. And, questions always exist without exception, and in the case of options, they can be divided into 3 or 4 depending on the level.

At this time, assuming that a question is selected for the test taker's level test, when the test taker gets the first question right, a more difficult question can be selected, and when the test taker gets the first question wrong, an easier question can be selected. . In this way, it is possible to test the candidate's ability by determining the next question according to the candidate's ability. The first question may be selected differently according to the age of the candidate or may be selected in consideration of the candidate's pre-stored ability. The selected questions may be stored in the data server 220 of the system 200 in the form of a question bank, and questions having a similar degree of similarity and difficulty may be selected and presented to the candidate. As such, in one embodiment of the present invention, previously presented questions may be analyzed and stored in the data server 220, and the questions stored in the data server 220 may be selected in real time and provided to the candidate. To this end, as shown in FIG. 3, a data server 220 in which numerous items and data analyzed for the items are stored, and a test taker terminal 210 that accesses the data server 220 and conducts an English test. The system 200 for the English proficiency test must be provided. The data server 220 classifies past questions into test categories, sections, parts, question numbers, options, correct answers, percent correct answers, distractors, keywords, themes, difficulty, Prompt Tense, Word Class, Intelligence Classification, Type, Communication Function, Question Types, Detail Questions Types, Inference, Prompts Structures , Prompt Types, Grammatical Points, etc., digitized and stored by group.

In one embodiment of the present invention, in order to analyze the similarity of items, the fingerprints of the previously performed items are classified according to a plurality of themes, the questions are classified according to a plurality of question types, and among the options, the answers are not correct. The reason for not answering the choice is classified into a plurality of types, and a predetermined score is given to an item belonging to the same group for each of the passage, question, and reason for not being an option, and then a predetermined score is given for the passage, question, and option. In case of obtaining the above, judge with similar questions. For example, the classification criteria for the fingerprint are Actions, Activities, Animals, Art, Careers, Clothing, Colors, Food, It may include at least one of Emotion, Entertainment, Health, Music, and Nature. The above questions are classified according to a number of question types, but the classification criteria for the question types are: finding an appropriate answer for a Who question, finding an appropriate response for a What or Which question, finding an appropriate response for a Where question, finding an appropriate response for a When question, and how It can include one or more of finding the right answer to the question.

In addition, the classification criteria for the reason for the non-correct answer of the non-correct option among the above options are Ambiguous pronouns, Confusing words, Keyword related expressions, Keyword unrelated expressions , Missing keywords, overlap, wrong description, wrong meaning, wrong sequence, and opposite meaning. there is. The ambiguous pronoun means a case in which an ambiguous pronoun that does not match the fingerprint is selected, and the confusing words mean a case in which a word similar to a key word in the fingerprint is selected but wrong, and an expression related to the key word (Keyword related) means a case in which an option is selected that is related to the key word or related word in the passage but does not match the overall flow, and Keyword unrelated (N/A)) is related to the key word in the passage It means the case of selecting an option with no words, Missing keywords means the case of selecting a key word or an option without a key word in the fingerprint, and the same expression (Overlap) is the same as the expression mentioned in the fingerprint Wrong description means choosing an option that is not appropriate for use, wrong description means choosing an option with an incorrect description of the text, and wrong meaning means choosing an option that misunderstands the meaning of the text. Wrong sequence means a case in which an option is selected in the wrong order in the flow of conversation, and a wrong situation means a case in which an option is selected in a wrong understanding of the context of the text. At this time, the similarity of the fingerprint, question, and option may be one basis for determining the similarity depending on whether the arrangement of sentences or phrases is arranged adjacent to each other. That is, for the similarity analysis, the text similarity is measured in a paragraph unit by combining the fingerprint and the text of the question, and the similarity is measured including the type of option. For example, when 'I' is 1 time in 'I love you', 'love' is 2 times, and 'you' is 3 times, I' is 1 time in 'I love her' and 'love' is 3 times. If number 2 and 'her' are set to number 4, these two sentences have the order of 1,2,3 and the order of 1,2,4, so they are highly likely to be judged as similar sentences. However, 'She loves me', if 'She' is 5 times, 'loves' is 2-1, and 'me' is 6 times, it has a layout of 5,2-1,6, so 1,2,3 It is highly likely to be judged as a sentence different from 'I love you', which has the order (vector) of As such, in one embodiment of the present invention, similarity is determined by considering words and word order for fingerprints, questions, and choices.

On the other hand, in one embodiment of the present invention, when the fingerprints are classified by core theme, 3 points are given if they are in the same group, 5 points are given if they are in the same group among the types of questions, and options that are not correct among the correct answers 2 points are given to those who are included in the same group for the reason why the answer is not correct, and the total points are sorted in the order of highest total points, and the items with the highest total points are selected and provided to the test taker. The score is only an example and is not limited thereto.

As such, in one embodiment of the present invention, fingerprints are classified for one question, questions are classified, and reasons for non-correct answers of non-correct options among options are classified and stored in the data server 220.

In addition, in an embodiment of the present invention, in order to predict the selection rate of each distractor, the selected test item and the selection rate for each distractor in each of the previously presented items are used as a dataset. If there are 4 options for each of the above questions and the correct answer is B in the tested test question, the correct answer rate for the question itself is expected to be 76%, but the selection rate for option A is 20% and the selection rate for option C is 3%, If the selection rate of option D is 2%, it may be different from the items with a selection rate of less than 10% among questions with a correct rate of about 76%. That is, the difficulty of the item can be judged by the correct answer rate, and furthermore, the difference between the selection rates for selecting an option other than the correct answer must be considered. For example, assuming that the rate of correct answers for two questions is 70%, the difficulty level when the selection rate of the wrong answer with the highest selection rate in the first question is 25% and the incorrect answer with the highest selection rate in the second question is 10%. , even if the percentage of correct answers is the same, it cannot be said that the difficulty is always the same.

In this way, after the data server 220 randomly selects an item (S150), when an item is selected in consideration of similarity and difficulty, the selected item is transmitted to the test taker terminal 210 (S160). The test taker solves the transmitted questions and continuously transmits them from the data server 220 to the test taker terminal 210, whereby the test taker's English proficiency test proceeds.

The above detailed description should not be construed as limiting in all respects and should be considered illustrative. The scope of the present invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

110: fingerprint, 120: question, 130: option, 200: system, 220: data server, 210: test taker terminal,

Claims (6)

In the method of selecting questions in the system for the English proficiency test,
Obtaining source data using a data server in which previously performed English questions are stored;
constructing a dataset by the data server by analyzing the source data into a plurality of pieces of data information including keywords, themes, difficulty levels, question formats, fingerprint structures, and fingerprint formats;
Grouping and classifying, by the data server, questions having similar problem types in the constructed dataset;
Grouping and classifying, by the data server, questions having similar difficulties in the constructed dataset;
The data server randomly selects an item in consideration of the classified similarity and difficulty level,
The English questions include a prompt that provides contextual understanding or background knowledge for solving the problem, a question that defines the nature of the question, including the content to be asked, and a distractor that includes the correct answer. include,
Classifying the fingerprints according to a plurality of themes, classifying the questions according to a plurality of question types, and classifying reasons why non-correct answers among the options are not correct answers into a plurality of types,
A predetermined score is given to questions belonging to the same group for each of the fingerprints, questions, and reasons for not being an option, and then judged as a similar item when a predetermined score or more is obtained for the fingerprint, question, and option. How to select questions for the English proficiency test. According to claim 1,
The classification criteria for the fingerprint are Actions, Activities, Animals, Art, Careers, Clothing, Colors, Food, and Emotion. ), entertainment, health, music, and nature. According to claim 2,
The above questions are classified according to a number of question types, but the classification criteria for the question types are: finding an appropriate answer for a Who question, finding an appropriate response for a What or Which question, finding an appropriate response for a Where question, finding an appropriate response for a When question, and how A question selection method for an English proficiency test comprising at least one of finding an appropriate answer to a question. According to claim 3,
Among the above options, the classification criteria for the reason why the non-correct answer is not correct are Ambiguous pronouns, Confusing words, Keyword related expressions, Keyword unrelated expressions, Keywords Characterized in that it includes at least one of missing keywords, identical expression (Overlap), wrong description (Wrong description), wrong meaning (Wrong meaning), wrong sequence (Wrong sequence), and Wrong situation (Wrong situation) How to select questions for the English proficiency test. According to claim 1,
A problem selection method for an English proficiency test, characterized in that when the fingerprint is provided as a picture, the picture is replaced with a description of information and pictures necessary for problem solving in sentences or text. According to claim 1,
The method for selecting a question for an English proficiency test, further comprising, after the data server randomly selecting an item, transmitting the selected item to the test taker terminal by the data server.

Images