TW201428666A - Method for evaluating learning outcomes of individual concept and computer readable media thereof - Google Patents

Abstract

The invention relates to a method for evaluating learning outcomes of an individual concept and a computer readable media thereof. The method comprises the following steps: making a test papers according to a composition of a standard concept, wherein each test question of the test papers includes at least one concept; identifying associated parameters of each two above mentioned concepts and the difficulty of each test question; considering the percentage of correct for answering the test question, the difficulty of the test question and the confidence of answering the question after the subject answers the test papers; taking calculation results of the above three values and their corresponding membership functions as input variables of a fuzzy inference operation to further obtain a learning outcomes of each individual concept. Accordingly, it can understand the actual learning ability of the subject more objectively and accurately by considering the percentage of correct for answering the test question and additionally considering the difficulty of the test question and if the subject gets the correct answer because of luck.

Description

Individual concept learning effectiveness evaluation method and computer readable media

The invention relates to an individual concept learning evaluation method and a computer readable medium thereof, in particular to a test result of a subject, and at the same time, three diagnostic factors for determining the correct rate of the answer, the difficulty of the test, and the confidence of the answer, The fuzzy theory is used to infer the learning effect of the learners, so as to improve the accuracy and objectivity of the results of the concept learning, and to solve the problems caused by the accuracy of the questions.

As everyone knows, in addition to the face-to-face classes and assignments in the school, students are required to participate in various types of examinations that are held regularly or irregularly for each subject (such as usual exams and periods). The senior high school entrance examination, the final exam, etc., to test the students' effectiveness in learning the subjects through the test results. For a long time, the teaching teachers only judged the students' good and bad standards by answering the correct answer rate. In addition to the correct rate, there are still many factors (such as the difficulty of the topic and whether the students are guessed by luck). It seems that they must also be taken into account, so how to correctly assess the learner's learning outcomes and objectively and accurately understand the students or The actual level of learning of the subject, in order to help the teacher to easily check the student's learning effectiveness and provide suitable test assessment materials, becomes an important and challenging subject; in addition, in the process of learning the selection of personnel, the most common The method used is to review the autobiographical resume of the job seeker for an interview or a professional examination, but in most cases only one time In the industry examination, if only the accuracy of the answer is the main criterion for judging the pros and cons of the job seeker's ability, and the difficulty of ignoring the question and whether the candidate is guessed by luck, the company will not be able to effectively find out what is needed. Professional.

Please refer to the Republic of China Patent No. M332233, "A Framework for Tracking Progress and Regression Status", which uses a computer to store multiple test scores through a network system and a web server, and with a statistical program each time the student The scores of the exams, and the progress or regression of the scores of the exams; the statistical procedures are the scores after passing the test, the PR value, the average, the standard deviation, the five standards (top mark, pre-mark, average mark, post mark, Bottom standard) Various statistical data analyze students' learning orientation in order to teach students in accordance with their aptitude. However, this kind of structure can only monotonously sort out the progress of students' scores and rankings. It is not only impossible to objectively and accurately understand the actual learning of students. Ability, and can not effectively tell students where the wrong concept is, or tell students that they should re-learn the part of the course, resulting in considerable efficiency and efficiency in learning.

Today, the inventor is not in a position to be objective and precise in the actual implementation of the above-mentioned methods for assessing a certain concept or learning ability. It is a tireless spirit, and through its rich professional knowledge and years of practical experience. The invention was assisted and improved, and the present invention was developed based on this.

The main purpose of the present invention is to provide an evaluation method for individual concept learning effectiveness, which considers the test results of the subject, considers the three diagnostic factors of the correct answer rate, the difficulty of the test question, and the confidence of the question, and uses the fuzzy theory to infer the learner's Learning effectiveness, in order to improve the accuracy and objectivity of the results of the concept of learning, solve the traditional problem of only the correct answer rate as a problem to judge the students' learning outcomes.

In order to achieve the above-mentioned implementation purposes, the inventors have developed the following implementation techniques, firstly based on a standard concept composition to produce a test paper, wherein each test question in the test paper contains at least one concept; then, according to the standard concept composition definition Two or two related parameters of the above concept; continuation, define the difficult parameters of each test, and after the test subject answers the test paper, consider the correct rate of the answer, the difficulty of the test, and the confidence of the answer, the calculation results of the above three Taking the corresponding attribution function as the input variable of the fuzzy inference operation, and then finding each individual concept Learning effectiveness; in addition to considering the correct rate of answering questions, it is also necessary to consider the difficulty of the question and whether the candidate is guessed by luck. This not only enables the instructor to understand the actual situation of the student or the subject more objectively and accurately. Learning ability, companies can also effectively find the professional talents that meet the needs.

In an embodiment of the present invention, the correct rate is calculated by dividing the sum of the weighted values in the concept by the sum of the weights of the concepts; and the difficulty can be divided by the sum of the difficulty of answering the questions in the concept divided by the concept. The sum of the difficulty of the test questions; the confidence factor is divided by the confidence parameter corrected by the concept according to the answer, divided by the sum of the confidence parameters of the correct answer, wherein if the concept X is the knowledge of the concept Y, when the concept Y is wrong, the concept X The corrected confidence parameter is the confidence parameter before the correction minus the first adjustment parameter. The first adjustment parameter is the difficulty parameter of the concept X and the concept Y multiplied by the difficult parameter of the test question, and when the concept X answer is correct, the concept X The corrected confidence parameter is the confidence parameter before the correction plus a second adjustment parameter, and the second adjustment parameter is the difficult parameter of the test question.

In an embodiment of the invention, the fuzzy method of the fuzzy inference operation may be selected from the group consisting of minimum-minimum-maximum, minimum-multiplication-maximum, multiplication-multiplication-maximum, minimum-multiplication-sum, or phase One of the multiplication-multiplication-sum method is preferably selected from the Mandani minimum-minimum-maximum operation method; and the fuzzy inference operation blur extraction method can be selected from the center of gravity defuzzification, area and gravity center solution. One of the fuzzification, the central defuzzification of the largest area, or the average defuzzification of the maximum value is preferably selected from the method of center of gravity defuzzification.

In addition, as described above, the method for evaluating individual concept learning effectiveness of the present invention may be applied to a hardware or a soft body, a suitable place or a combination thereof, and thus some aspects or portions thereof may be embedded in a computer readable medium. The executable instruction (i.e., the code) is such that when the code is loaded and executed by a machine (e.g., a computer), the machine becomes a means for performing the method of evaluating the individual concept learning effectiveness of the present invention.

(S1)‧‧‧Step one

(S2)‧‧‧Step 2

(S3) ‧ ‧ Step 3

(S4)‧‧‧Step four

(S5) ‧ ‧ step five

First Figure: Flowchart of the steps of a preferred embodiment of the present invention

The second figure: the fourth embodiment of the specific practical embodiment of the present invention

Third figure: a attribution function used to blur the correct rate of a specific practical embodiment of the present invention

Fourth figure: a attribution function used to obfuscate difficulty in a specific practical embodiment of the present invention

Fifth figure: a attribution function used to obfusce confidence in a practical embodiment of the present invention

Figure 6: A attribution function used to obfuscate the learning effect of a specific practical embodiment of the present invention

The object of the present invention and its structural design and advantages will be explained in the light of the preferred embodiments shown in the following drawings, so that the reviewing committee can have a more in-depth and specific understanding of the present invention.

First, in order to better understand the present invention, the basic concept of concept mapping will be briefly explained first, and conceptual composition can also be called conceptual composition; conceptual composition is an image used by learners to understand the relationship between concepts. Representation is mainly composed of "concepts" and "conjunctions". The two concepts are connected by lines, and the lines are marked by the conjunctions to describe the relationship between concepts. Therefore, the simplest concept composition is composed of two The concept consists of a concept, such as: grass - is - green, "grass" and "green" represent two different concepts, respectively, "Yes" means connecting the connection between the two concepts, through the concept Meaningful connections, and the concept of effective hierarchical structure, so that the original scattered concept can be re-integrated into a visually well-organized learning textbook; in addition to helping to connect general concepts and specific concepts, it can also add relevant concepts. To connect new information, the most powerful link is to connect previously dispersed chapters or topics, and unite them together to form a complete conceptual structure; The four arithmetics of basic mathematics include the four concepts of "addition", "subtraction", "multiplication" and "division", and the relationship between these concepts is gradual, meaning that before multiplication and division, Yes It is necessary to complete the addition and subtraction learning, that is to say, the concept of "multiplication" or "division" is the backup knowledge of the concept "addition" and the concept "subtraction". Therefore, the learning must be preceded by "addition" and "subtraction". After having a certain understanding of the knowledge, we can carry out the following "multiplication" or "division" learning; from the above we can understand that conceptual composition is a visual image representation, it seems that every concept is There is no connection, but through the combination of the connection between ideas and concepts, the network of concept representations can be completely presented, so that the connection materials are refined from point to line to face. The learner can better grasp the focus of the learning materials; then, referring to the first figure, the flow chart of the preferred embodiment of the method for evaluating the individual concept learning effectiveness of the present invention includes the following steps:

Step 1 (S1): According to a standard concept composition, a test paper is prepared, wherein each test question in the test paper contains at least one concept; the standard concept composition can be based on the curriculum standard of the Ministry of Education (for example, the national, medium and small nine The annual syllabus, etc.) is drawn or produced by experts in the field; for example, the distribution of the questions Q ₁ ~Q _m and each concept C ₁ ~C _p can be expressed as the following table, Q _t C _i =1 indicates that the test question is related to the concept, and Q _t C _i =0 indicates that the test question is not related to the concept, wherein 1≦t≦m, 1≦i≦p;

Step 2 (S2): According to the standard concept composition, define the associated parameters of the above two concepts: for example, the distribution parameters of the two concepts may be expressed as shown in the following table, where W _xy represents the concept C _x and the concept C _y Associated parameters, and the concept C _x is the concept of the concept of C _y , and 0 ≦ W _xy ≦ 1;

Step 3 (S3): define the difficult parameters of each test question, wherein the test questions Q ₁ ~Q _m and the corresponding difficult parameter D _t can be as follows, wherein 0 ≦ D _t ≦ 1, and 1 ≦ t ≦ m; It is worth noting that the associated parameter (W _xy ) of step two (S2) and the difficult parameter D _{t of} step three (S3) may be defined by expert experience in the field, or may be an algorithm and software tool with expert rule of thumb. Produced

Step 4 (S4): After the test subject answers the test paper, consider the correct rate of the answer, the difficulty of the test, and the confidence of the answer, and calculate the correct rate of each question, the difficulty of the test, and the confidence of the test. Wherein, the correct answer rate can be calculated by dividing the sum of the weights in the concept by the sum of all the weights of the concept, meaning Where ac _ij represents the correct rate of the S _j subjects for the concept C _i , where 1≦j≦n, 1≦t≦m, and 1≦i≦p; further, R _tj represents the S _jth The respondent's answer status to the test Q _t , when R _tj =1, indicates that the S _j subjects answered correctly to the test Q _t , and when R _tj =0, the S _j test pairs are indicated. The question Q _t answers incorrectly; in addition, the difficulty of the test can be divided by the sum of the difficulty of answering the test questions in the concept divided by the difficulty of the test questions covered by the concept, that is, Where D _ij represents the difficulty of the S _j subjects for the concept C _i ; finally, the confidence of the answer is divided by the confidence parameter corrected by the concept according to the answer, divided by the sum of the confidence parameters of the correct answer, and if the concept X For the concept Y backup knowledge, when the concept Y answer is wrong, the confidence parameter after the concept X is corrected is the confidence parameter before the correction minus the first adjustment parameter, and the first adjustment parameter is the correlation parameter of the concept X and the concept Y. The difficult parameters of the test questions, and when the concept X answer is correct, the confidence parameter after the concept X correction is the confidence parameter before the correction plus a second adjustment parameter, and the second adjustment parameter is the difficult parameter of the test question, that is, Where CL _ij represents the average confidence that the S _j subjects answered correctly for the concept C _i , and ;among them, , , Corrected the confidence parameter for the idea node, and , , The confidence parameter before the correction of the concept node;

Step 5 (S5): The calculated correctness rate, difficulty degree, and confidence are calculated as the input variables of the fuzzy inference operation by using the corresponding membership function as the input variable of the fuzzy inference operation. Learning acgmentment; wherein the fuzzification method of fuzzy inference operation is selected from Min-Min-Max, Min-Product-Max, and multiplication - Product-Product-Max, Min-Product-Sum or Product-Product-Sum method, preferably selected From the Mandani Min-Min-Max method; the fuzzy inference method can be selected from the Center of Gravity, the area and the center of gravity to blur ( Center of Sum), one of the Center of Maximum of the largest area or the Center of Mean of the maximum, preferably selected from the center of gravity defuzzification method.

Next, the method for evaluating the individual concept learning effectiveness of the present invention is implemented as a specific embodiment under the computer system of a processor, and the specific practical example of the present invention further proves that the method for evaluating individual concept learning effectiveness of the present invention can be practical. The scope of the application is not intended to limit the scope of the invention in any way. First, please refer to the second figure, which is a schematic diagram of the four standard operation standard diagrams of the specific practical embodiment of the present invention, wherein C1~C4 respectively indicate The concepts of addition, subtraction, multiplication and division, and the value between the two concepts is its associated parameter. For example, the correlation parameter between the addition concept C1 and the multiplication concept C3 is 0.7, so the distribution relationship of the related parameters of the following two concepts can be obtained. table:

Then, according to the above standard concept composition, a test paper is prepared, and the difficult parameters of each test question are defined by the expert experience in the field, wherein each test question in the test paper contains at least one concept; finally, at the first After the test subject (S1) answers the test paper, the correctness of each question answer status (R _tj ) is recorded. When R _tj =1, the test answer is correct, and when R _tj =0, the answer is Error; as shown in the following table, the relationship between the questions Q ₁ ~ Q ₅ relative to each concept C ₁ ~ C ₄ , difficult parameters and the status of the answer.

Therefore, by the above formulas of correctness, difficulty, and confidence, the following can be obtained separately; for the correct rate: For the difficulty, the calculation is as follows: Finally, for confidence: Then, taking the additive concept C1 as an example, the correctness rate, difficulty degree, and confidence calculated by the above calculation are respectively blurred by the corresponding attribution function, and the attribution functions are respectively defined in the third to fifth graphs, and the correct rate is The difficulty and confidence parameters are defined in the range of 0 to 100, and have three parameter settings: low attribution function (low), medium attribution function (medium), and high attribution function (high); There are many kinds of custom attribution functions, including triangle functions, trapezoidal functions, inverted clock functions, Gaussian functions, etc. This embodiment uses a trapezoidal function as an example, but is not limited; and the addition concept C1 fuzzification process As described below: the correctness (100%) of the addition concept C1 is blurred by the attribution function defined in the third figure, and the degree of low accuracy is 0, and the degree of moderate accuracy is 0, and belongs to The degree of high degree of accuracy is 1; then the difficulty degree (100%) of the addition concept C1 is blurred by the attribution function as defined in the fourth figure, and the degree of low difficulty is 0, which is the degree of difficulty. It is 0, and the degree of high difficulty is 1; finally, the confidence (33.8%) of the additive concept C1 is blurred by the attribution function of the fourth graph, and the degree of low confidence is 0.62, belonging to the middle. The degree of confidence is 0.38, and the degree of high confidence is 0. Then, a corresponding learning effect output value can be obtained by, for example, a fuzzy rule of learning efficiency defined by an expert in the field as described below. The fuzzy inference operation of the present embodiment is fuzzified according to the theory of Mamdani Min-Min-Max, but is not limited, and other fuzzification methods such as Min-Prod may be used. The uct-Max, Product-Product-Max, Min-Product-Sum or Product-Product-Sum method is implemented, and the rule has a total of 3 ³ = 27, and 5 of them are listed here to avoid excessive length; Therefore, taking the rule R1 as an example, the degree of accuracy is 0, the degree of difficulty is 0, and the degree of confidence is 0.62, so min(0, 0, 0.62) = 0, so The degree of learning effectiveness belongs to Very poor is 0; R25 is an example. The degree of accuracy is high. 1. The degree of difficulty is high, and the degree of confidence is low is 0.62. Therefore, min(1,1,0.62)=0.62, so the learning result belongs to Good. The degree is 0.62; the case of R26 is that the correct rate is high, the degree of difficulty is high, the degree of confidence is high, and the degree of confidence is 0.38, so min(1,1,0.38)=0.38, Therefore, the degree of learning achievement is Good, which is 0.38. In addition, the results of all the rules are accumulated to obtain the degree of learning effectiveness corresponding to each fuzzy situation as follows:

Finally, the learning effect is blurred by the attribution function as defined in the fifth figure, wherein the learning effect parameter defines the range from 0 to 100, and has poor (Poly poor), Poor, medium, and good. (Good), with 5 parameters such as the good function of Very good, and the output is quantized into five discrete outputs (60, 70, 80, 90, 100) by the center of gravity defuzzification method. The learning effect of the addition concept C1 is: It is worth noting that the above-mentioned solution fuzzy method is the focus of the fuzzy method, but it is not limited, and other defuzzification methods such as the area and the center of gravity can be used to blur the maximum area. Unfuzzification or maximum flatness The average de-fuzzification method is implemented; thereby, the individual concept learning effectiveness evaluation method of the present invention calculates each diagnostic factor (correctness rate, test difficulty, confidence) based on the test result of the subject, and then uses the fuzzy theory. It is inferred that the testee's learning effect is compared with the traditional only correct answer rate as the criterion for judging the student's learning performance or selecting the merits of the candidate. In addition to considering the correct answer rate, the present invention also considers the difficulty of the problem and the test. Whether the person is lucky or not, not only enables the instructor to understand the actual learning ability of the student or the subject more objectively and accurately, but also enables the company to effectively find the professional who meets the needs.

In addition, as described above, the method for evaluating individual concept learning effectiveness of the present invention may be applied to hardware or software, where appropriate, or a combination thereof, and thus some aspects or portions thereof may be readable by a computer. Executable instructions (ie, code) in the media, wherein the computer readable recording medium may include the following devices: an electrical transmission module (electronic medium) having one or more wires, and a portable computer magnetic disk ( Magnetic media), random access memory (RAM) (electronic media), read-only memory (ROM) (electronic media), erasable programmable read-only memory (EPROM, EEPROM, Flash memory) (electronic Medium), optical fiber (optical medium), and a portable compact disc read-only memory (CDROM) (optical medium), etc., when the code is used by the machine (example) When a computer is loaded and executed, the machine becomes a device for performing the individual concept learning evaluation method of the present invention, and in the case where the code is executed in a programmable computer, the computer device usually includes a process. A processor-readable storage medium (including volatile or non-volatile memory and/or storage elements), at least one input device, and at least one output device.

In summary, the method for evaluating individual concept learning effectiveness of the present invention can achieve the intended use efficiency by the above-disclosed embodiments, and the present invention has not been disclosed before the application, and has completely complied with the patent law. Regulations and requirements. 爰Issuing an application for a patent for invention in accordance with the law, and asking for a review, and granting a patent, is truly sensible.

The illustrations and descriptions of the present invention are merely preferred embodiments of the present invention, and are not intended to limit the scope of the present invention; those skilled in the art, which are characterized by the scope of the present invention, Equivalent variations or modifications are considered to be within the scope of the design of the invention.

(S1)‧‧‧Step one

(S2)‧‧‧Step 2

(S3) ‧ ‧ Step 3

(S4)‧‧‧Step four

(S5) ‧ ‧ step five

Claims (10)

An individual concept learning evaluation method includes the following steps: Step 1: preparing a test paper according to a standard concept composition, wherein each test question in the test paper contains at least one concept; step two: according to the The standard concept composition defines the associated parameters of the two concepts; step 3: define the difficult parameters of each test question; step 4: after the test subject answers the test paper, calculate the correct rate and difficulty level corresponding to each concept. And the confidence level; and step 5: the correctness rate, the difficulty degree, and the attribution function corresponding to the confidence degree are used as input variables of the fuzzy inference operation, thereby obtaining the learning effect of each individual concept. For example, the individual concept learning effectiveness evaluation method described in the first paragraph of the patent application scope, wherein the correctness rate of the concept is the sum of the weighted values of the correctness of the concept divided by the total weight value of the concept. For example, the individual concept learning effectiveness evaluation method described in claim 1 of the patent scope, wherein the difficulty of the concept is the sum of the difficulty of answering the test questions in the concept divided by the difficulty of the test questions covered by the concept. For example, the individual concept learning effectiveness evaluation method described in item 1 of the patent application scope, wherein the confidence of the concept is determined by the concept The confidence parameter corrected by the problem situation is divided by the sum of the confidence parameters of the concept. For example, the individual concept learning effectiveness evaluation method described in item 4 of the patent application scope, wherein if the concept X is the knowledge of the concept Y, when the concept Y is wrong, the confidence parameter after the concept X is corrected is the confidence parameter before the correction. Go to a first adjustment parameter, the first adjustment parameter is the difficulty parameter of the concept X and the concept Y multiplied by the difficult parameter of the test question, and when the concept X answer is correct, the confidence parameter of the concept X is the confidence before the correction The parameter is added with a second adjustment parameter, which is a difficult parameter of the test question. For example, the method for evaluating individual concept learning effectiveness described in claim 1 is wherein the fuzzy inference method is selected from the group consisting of min-min-max, min-multiply-maximum, multiply-multiply-maximum, One of the minimum-multiply-sum or multiply-multiply-sum method. For example, the method for evaluating individual concept learning effectiveness described in claim 6 of the patent application scope, wherein the fuzzy inference method of the fuzzy inference operation is selected from Mandanning's minimum-minimum-maximum. For example, the method for evaluating individual concept learning effects described in claim 1 of the patent scope, wherein the fuzzy inference method of the fuzzy inference operation is selected from the center of gravity defuzzification, the area and the center of gravity solution blurring, The center of the largest area is defuzzified or the average of the maximum is defuzzified. For example, the method for evaluating individual concept learning effectiveness described in claim 8 of the patent scope, wherein the fuzzy inference method of the fuzzy inference operation is selected from the method of center of gravity defuzzification. A computer readable medium comprising at least a computer executable module, the computer executable module having computer executable instructions, when a computer executes the computer executable instructions, executable as claimed in items 1 to 9 The method of any of the preceding claims.