A method and system for constructing evaluator's cognitive structure map using comments

Abstract

The present invention discloses a method for constructing an evaluator's cognitive structure map using comments, including: structurally encoding the evaluator's comments according to the evaluation criteria and knowledge system; visualizing the encoding results as the evaluator's cognitive structure map through mathematical statistics and weight calculation of nodes and lines; analyzing the corresponding knowledge level, level and structural characteristics according to the evaluator's cognitive structure map, so as to provide a powerful decision-making reference for improving teaching and learning and implementing precision teaching. The application objects determined by the present invention based on the cognitive framework theory are: the comments and texts generated by learners who have mastered similar knowledge and skills in a certain professional field when participating in professional evaluation activities. The present invention also includes a system for implementing the above-mentioned method of constructing an evaluator's cognitive structure map using comments.

Description

Technical Field

[0001] The present invention belongs to the field of educational evaluation in pedagogy, relates to data statistics and analysis technology, and is a method and system for constructing a cognitive structure map of an evaluator using comments. Background Art

[0002] In the process of teaching, teachers need to understand the teaching effect and control the teaching quality through formative evaluation methods such as correcting students' homework. However, with the popularity of multiple evaluation methods, educational evaluation has also begun to shift from focusing on examinations to promoting learners' learning. More and more scholars have carried out analysis of comments, texts and other data generated in teaching activities in order to implement a comprehensive evaluation of learners. At present, the main analysis methods for comments are content analysis and natural language processing.

[0003] Ye Ping used mathematical statistics and a combination of qualitative and quantitative analysis methods to analyze the types of focus in the collected comments and compare the results of two speeches, proving that peer evaluation can significantly improve learners' English speaking scores; Zhang Yibing divided the content of students' comments on the online platform into multiple dimensions through sub-item scoring, text analysis and content analysis, and conducted statistical analysis on the content of peer evaluation. He used a line graph to describe the changing trend of the content of peer evaluation, and then judged the role of peer evaluation in promoting student learning; Dong Jiali used mathematical statistics to analyze the distribution relationship between the three dimensions of English writing metacognition in the comments; Li Hua used mathematical statistics to conduct statistical analysis on the content of peer evaluation posted by students on the learning platform, showing the changes in their views, and then analyzed the impact of essay peer evaluation on students' writing level.

[0004] Most of the existing comment analysis methods set different dimensions and then analyze the performance of comments in each dimension to draw conclusions related to mutual evaluation. Lack of mining of learner dynamic information data in comments makes it impossible to test the consistency between learners' cognitive structure level and their social practice, and even more impossible to reveal the differences in cognitive structure and level of comments written by learners with different grades from the perspective of educators. It is difficult to help educators obtain information about the evaluator's thinking mode and cognitive structure from the comments, and then improve and implement subsequent teaching activities in a targeted manner. Summary of the invention

[0005] The present invention aims to solve the above-mentioned deficiencies in the prior art and provides a method and system for constructing an evaluator's cognitive structure map using comments.

[0006] The present invention analyzes the knowledge level, level and structural characteristics of the evaluators corresponding to different comments through learning analysis techniques such as cognitive structure analysis methods and mathematical statistics.

[0007] A method for constructing an evaluator's cognitive structure map using comments, comprising:

[0008] 1) Structural coding of the evaluator's comments based on the evaluation criteria and knowledge system;

[0009] 2) Visualize the encoding results into cognitive structure maps through mathematical statistics and weight calculation of nodes and lines;

[0010] 3) Analyze the corresponding knowledge level, level and structural characteristics according to the evaluator's cognitive structure map to provide a powerful decision-making reference for improving teaching and learning and implementing precision teaching.

[0011] Further, the step 1) specifically includes:

[0012] 1.1) Review collection: Collect reviews written by evaluators and remove invalid reviews;

[0013] 1.2) Determine the dimensions: Construct the cognitive framework of a certain profession or occupation of the evaluator based on the knowledge and cognition that the evaluator should have, and divide the dimensions of knowledge elements as the basis for comment coding; that is, define the knowledge elements of the cognitive framework of a certain profession or occupation of the evaluator, from f1…f n Composition, where f represents a specific coded element in a specified time window, which is a specific form of some skills, knowledge, values, identity or epistemology, and n dimensions are usually used to represent n knowledge elements;

[0014] 1.3) Comment encoding: Use binary values ​​"0" and "1" to encode the text in the comment to determine whether a comment contains the above dimensions; if the comment contains words or phrases representing the above dimensions, mark "1" on the dimension, and mark "0" on the dimension not involved, and finally convert each comment into a set of "1" and "0" codes of the dimension.

[0015] Preferably, the step 2) specifically includes:

[0016] 2.1) Data processing: Each comment of each evaluator is set as a section unit. The comment statements of all dimensions in a section unit are for the same person being evaluated, so these dimensions are interrelated; the dimensions between different sections are not for the same person being evaluated, so they are not related; in a section unit, which dimensions appear in the dimension coding set and appear at the same time (expressed as the coding of related dimensions is "1"), which reflects the evaluator's understanding of the knowledge elements and the corresponding cognitive thinking when evaluating;

[0017] All comments of a certain evaluator are grouped into a group unit, in which there are different comments of the evaluator on different evaluated persons; the union of the dimension coding sets of all the sub-units in a group unit comprehensively reflects the knowledge level of the evaluator in the framework of professional or occupational knowledge, as well as his cognition of such knowledge;

[0018] Create a matrix of the group units corresponding to each coded evaluator to quantify the correlation between the size and dimension of each code; suppose an evaluator wrote m reviews and defined the cognitive framework of a certain profession or occupation as n knowledge elements f1…f n Composition, where f represents a specific coding element, the following matrix A can be listed; where a mn It represents the coding value of the nth dimension in the mth review written by the evaluator. If its value is "1", it means that the knowledge element of the nth dimension is reflected in the mth review. If its value is "0", it means that the knowledge element of the nth dimension is not reflected in the mth review. The rows of the matrix represent the coding of the same review in each dimension, and the columns of the matrix represent the coding of a certain dimension in different reviews.

[0019] The coded data of each group unit (matrix) is displayed as a cognitive structure map in the form of a weighted network of nodes and lines: nodes represent dimensions, i.e., knowledge elements; the size of the node represents its weight, reflecting the number of times the knowledge element appears in the comments of the group unit; in the same node unit (row of the matrix), dimensions coded as "1" are colinear with each other, indicating that these dimensions are related in the evaluator's cognition; related dimensions are connected by lines, and the thickness of the lines represents their weight, reflecting the number of times the related dimensions appear together in the nodes of a group unit;

[0020] In this way, the cognitive structure map of the evaluator is constructed; each evaluator has his own unique cognitive structure map, which represents his understanding of knowledge elements and the correlation frequency of these knowledge elements, and further reflects his understanding of professional or occupational knowledge and the corresponding cognitive structure;

[0021] ①Node weight calculation: D i Represents knowledge element f i The weight of D i Equal to f of matrix A i The total number of times "1" appears in the column, that is

[0022] D i =a 1i +a 2i +a 3i +a 4i +…+a mi Formula 1

[0023] ② Connection weight calculation: Let fi , f j There are two knowledge elements, F i,j Represents knowledge element f i and f j The total number of associations in a group unit, Indicates that in the kth (k = 1, 2, ..., m) section unit, the knowledge element f i and f j Is it related?

[0024]

[0025]

[0026] 2.2) Data visualization: Draw a cognitive structure map of corresponding proportions based on the weight values ​​of nodes and lines.

[0027] Preferably, cognitive structure analysis is performed after data visualization: the coded data of all evaluators are visualized to obtain a cognitive structure map of each evaluator; in order to compare the differences in cognitive structures between different evaluators, their cognitive structure maps are subjected to operations such as reduction to more clearly reflect the differences in their knowledge levels and cognitive levels.

[0028] A system for implementing the method of constructing an evaluator's cognitive structure map using comments of the present invention comprises an evaluator's comment encoding module, an encoding result visualization module, and an evaluator's cognitive structure analysis module connected in sequence;

[0029] The evaluator comment coding module performs structured coding on the evaluator's comments according to the evaluation criteria and knowledge system; specifically, it includes the comment collection submodule, dimension determination submodule, and comment coding submodule which are connected in sequence;

[0030] The review collection submodule collects the reviews written by the reviewers and removes invalid reviews;

[0031] The dimension determination submodule constructs the cognitive framework of a certain profession or occupation of the evaluator according to the knowledge and cognition that the evaluator should have, and divides the dimensions of knowledge elements as the basis for comment coding; that is, the knowledge elements of the cognitive framework of a certain profession or occupation to which the evaluator belongs are defined, which are composed of f1…f n Composition, where f represents a specific coded element in a specified time window, which is a specific form of some skills, knowledge, values, identity or epistemology, and n dimensions are usually used to represent n knowledge elements;

[0032] The comment encoding submodule encodes the text in the comment with binary values ​​"0" and "1" to determine whether a comment contains the above dimensions; if there are words or phrases representing the above dimensions in the comment, the dimension is marked as "1", and the dimension not involved is marked as "0", and finally each comment is converted into a set of "1" and "0" codes of the dimension;

[0033] The coding result visualization module visualizes the coding result into a cognitive structure map through mathematical statistics and weight calculation of nodes and lines, including a data processing submodule and a data visualization submodule connected in sequence;

[0034] The data processing submodule sets each comment of each evaluator as a section unit. The comment statements of all dimensions in a section unit are for the same person being evaluated, so these dimensions are interrelated; the dimensions between different sections are not for the same person being evaluated, so they are not related; in a section unit, which dimensions appear in the dimension coding set and appear at the same time (expressed as the coding of related dimensions are all "1"), which reflects the evaluator's understanding of the knowledge elements and the corresponding cognitive thinking when evaluating;

[0035] All comments of a certain evaluator are grouped into a group unit, in which there are different comments of the evaluator on different evaluated persons; the union of the dimension coding sets of all the sub-units in a group unit comprehensively reflects the knowledge level of the evaluator in the framework of professional or occupational knowledge, as well as his cognition of such knowledge;

[0036] Create a matrix of the group units corresponding to each coded evaluator to quantify the correlation between the size and dimension of each code; suppose an evaluator wrote m reviews and defined the cognitive framework of a certain profession or occupation as n knowledge elements f1…f n Composition, where f represents a specific coding element, the following matrix A can be listed; where a mn It represents the coding value of the nth dimension in the mth review written by the evaluator. If its value is "1", it means that the knowledge element of the nth dimension is reflected in the mth review. If its value is "0", it means that the knowledge element of the nth dimension is not reflected in the mth review. The rows of the matrix represent the coding of the same review in each dimension, and the columns of the matrix represent the coding of a certain dimension in different reviews.

[0037] The coded data of each group unit (matrix) is displayed as a cognitive structure map in the form of a weighted network of nodes and lines: nodes represent dimensions, i.e., knowledge elements; the size of the node represents its weight, reflecting the number of times the knowledge element appears in the comments of the group unit; in the same node unit (row of the matrix), dimensions coded as "1" are colinear with each other, indicating that these dimensions are related in the evaluator's cognition; related dimensions are connected by lines, and the thickness of the lines represents their weight, reflecting the number of times the related dimensions appear together in the nodes of a group unit;

[0038] In this way, the cognitive structure map of the evaluator is constructed; each evaluator has his own unique cognitive structure map, which represents his understanding of knowledge elements and the correlation frequency of these knowledge elements, and further reflects his understanding of professional or occupational knowledge and the corresponding cognitive structure;

[0039] ①Node weight calculation: D i Represents knowledge element f i The weight of D i Equal to f of matrix A i The total number of times "1" appears in the column, that is,

[0040] D i =a 1i +a 2i +a 3i +a 4i +…+a mi Formula 1

[0041] ② Connection weight calculation: Let f i , f j There are two knowledge elements, F i,j Represents knowledge element f i and f j The total number of associations in a group unit, Indicates that in the kth (k = 1, 2, ..., m) section unit, the knowledge element f i and f j Is it related?

[0042]

[0043]

[0044] The data visualization submodule draws a cognitive structure map of corresponding proportions according to the weight values ​​of nodes and lines;

[0045] The evaluator's cognitive structure analysis module analyzes the corresponding knowledge level, level and structural characteristics according to the evaluator's cognitive structure map, providing a powerful decision-making reference for improving teaching and learning and implementing precise teaching.

[0046] The advantages of the present invention are: the proposed method of constructing an evaluator's cognitive structure map using comments and its system can reflect the evaluator's knowledge level and cognitive structure at a deeper level by drawing the evaluator's cognitive structure map, so as to help education implementers better understand the cognitive structure characteristics of learners and improve their teaching strategies in a targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flow chart of the method of the present invention;

[0048] Figure 2 This is an example diagram of the cognitive structure map of evaluator A generated using the present invention.

[0049] Figure 3 This is an example diagram of the cognitive structure map of evaluator B generated using the present invention. DETAILED DESCRIPTION

[0050] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0051] Figure 1 FIG. 4 is a flow chart showing a method for constructing an evaluator's cognitive structure map using comments in the present invention. Figure 1 As shown, the method includes:

[0052] Step 1: Review collection: Collect reviews written by evaluators and remove invalid reviews.

[0053] Step 2: Determine the dimensions: Construct the cognitive framework of a certain profession or occupation of the evaluator based on the knowledge and cognition that the evaluator should have, and divide the dimensions of knowledge elements as the basis for comment coding. That is, define the knowledge elements of the cognitive framework of a certain profession or occupation of the evaluator, from f1…f n Composition, where f represents a specific coded element in a specified time window, which is a specific form of some skills, knowledge, values, identity or epistemology, and n dimensions are usually used to represent n knowledge elements.

[0054] Step 3, comment coding: Use binary values ​​"0" and "1" to encode the text in the comment to determine whether a comment contains the above dimensions. If there are words or phrases in the comment that represent the above dimensions, mark "1" on the dimension, and mark "0" on the dimension that is not involved. Finally, each comment is converted into a set of "1" and "0" codes for the dimension.

[0055] Step 4, data processing: Set each comment of each evaluator as a section unit. The comment statements of all dimensions in a section unit are for the same person being evaluated, so these dimensions are interrelated; the dimensions between different sections are not for the same person being evaluated, so they are not related. In a section unit, which dimensions appear in the dimension coding set and appear at the same time (expressed as the coding of related dimensions is "1") reflects the evaluator's understanding of the knowledge elements and the corresponding cognitive thinking when evaluating.

[0056] All comments of an evaluator are grouped into a group unit, in which there are different comments of the evaluator on different evaluated persons. The union of the dimension coding sets of all the sub-units in a group unit fully reflects the evaluator's knowledge level in the framework of professional or occupational knowledge, as well as his cognition of such knowledge.

[0057] Create a matrix of the group units corresponding to each evaluator, and quantify the correlation between the size and dimension of each code. Assume that an evaluator has written m reviews, and the cognitive framework defining a certain profession or occupation consists of n knowledge elements f1…f n Composed of, where f represents a specific coding element, the following matrix A can be listed. mn It represents the coding value of the nth dimension in the mth review written by the evaluator. If its value is "1", it means that the knowledge element of the nth dimension is reflected in the mth review. If its value is "0", it means that the knowledge element of the nth dimension is not reflected in the mth review. The rows of the matrix represent the coding of the same review in each dimension, and the columns of the matrix represent the coding of a certain dimension in different reviews.

[0058] The coded data of each group unit (matrix) is displayed as a cognitive structure map in the form of a weighted network of nodes and lines: nodes represent dimensions, i.e., knowledge elements; the size of the node represents its weight, reflecting the number of times the knowledge element appears in the comments of the group unit; in the same node unit (row of the matrix), dimensions coded as "1" are collinear with each other, indicating that these dimensions are related in the evaluator's cognition. Related dimensions are connected by lines, and the thickness of the lines represents their weight, reflecting the number of times the related dimensions appear together in the nodes of a group unit.

[0059] In this way, the cognitive structure map of the evaluator can be constructed. Each evaluator has his own unique cognitive structure map, which represents his understanding of knowledge elements and the correlation frequency of these knowledge elements, and further reflects his understanding of professional or occupational knowledge and the corresponding cognitive structure.

[0060]

[0061] ①Node weight calculation: D i Represents knowledge element fi The weight of D i Equal to f of matrix A i The total number of times "1" appears in the column, that is

[0062] D i =a 1i +a 2i +a 3i +a 4i +…+a mi Formula 1

[0063] ② Connection weight calculation: Let f i , f j There are two knowledge elements, F i,j Represents knowledge element f i and f j The total number of associations in a group unit, Indicates that in the kth (k = 1, 2, ..., m) section unit, the knowledge element f i and f j Whether it is associated.

[0064]

[0065]

[0066] Step 5: Data visualization: Draw a cognitive structure map of corresponding proportions based on the weight values ​​of nodes and lines.

[0067] Step 6: Cognitive structure map analysis: Visualize the coded data of all evaluators to obtain the cognitive structure map of each evaluator. In order to compare the differences in cognitive structures between different evaluators, their cognitive structure maps can be subtracted to more clearly reflect their differences in cognitive structure.

[0068] The above-mentioned scheme proposed by the present invention is explained below according to specific examples:

[0069] In a peer review activity, five comments written by two evaluators A and B were collected. It is necessary to analyze the differences in the cognitive structures of evaluators A and B in a certain professional field from three dimensions. Through coding, the group unit matrix A of evaluator A and the group unit matrix B of evaluator B can be obtained as follows:

[0070]

[0071]

[0072] Node weight calculation: D i Represents the total number of times dimension i appears in the matrix

[0073] D i =a 1i +a 2i +a 3i +a 4i +…+a mi Formula 1

[0074] In the group cell matrix A:

[0075] D1=1+1+0+1+1=4

[0076] D2=0+1+1+1+0=3

[0077] D3=0+0+1+1+0=2

[0078] In the group cell matrix B:

[0079] D1=1+1+1+0+0=3

[0080] D2=0+1+0+1+1=3

[0081] D3=1+1+1+0+1=4

[0082] Connection weight calculation: F i,j represents the total number of associations between dimension i and dimension j in the matrix (association means dimensions that appear simultaneously in the same comment), Indicates the association between dimension i and dimension j in the kth section.

[0083]

[0084]

[0085] In the group cell matrix A:

[0086] F 1,2 =0+1+0+1+0=2

[0087] F 1,3 =0+0+0+1+0=1

[0088] F 2,3 =0+0+1+1+0=2

[0089] In the group cell matrix B:

[0090] F 1,2 =0+1+0+0+0=1

[0091] F 1,3 =1+1+1+0+0=3

[0092] F 2,3 =0+1+0+0+1=2

[0093] like Figure 2 (corresponding to evaluator A) and Figure 3 As shown in (corresponding evaluator B), according to the calculated ratio between node weights and the ratio between line weights, the corresponding cognitive structure map can be drawn (the ratio used this time is: the ratio of node areas is the square ratio of corresponding node weights, and the ratio of line thickness is the ratio of line weight sizes).

[0094] By comparing the cognitive structure maps of evaluators A and B, we can see that in terms of the frequency of use of knowledge elements: compared with evaluator B, evaluator A uses the knowledge elements corresponding to dimension 1 more frequently and has a more thorough understanding; while the frequency of use and degree of understanding of the knowledge elements corresponding to dimensions 2 and 3 are not as good as those of evaluator B.

[0095] In terms of the degree of association between knowledge elements: compared with evaluator B, evaluator A has a stronger connection between dimension 1 and dimension 2, which means that the association between the knowledge elements corresponding to dimension 1 and the knowledge elements corresponding to dimension 2 has a higher weight in his cognitive structure. The relevant knowledge elements are repeatedly used by evaluator A in the evaluation, while other knowledge elements are rarely used simultaneously.

[0096] This invention collects comments, performs structured coding on the comments of the evaluators according to the evaluation criteria and knowledge system, and visualizes the coding results as cognitive structure maps through mathematical statistics and weight calculation of nodes and lines, and then analyzes the corresponding cognitive levels, levels and structural characteristics. Finally, an experimental verification was carried out, and the results showed that the comment analysis method based on cognitive structure maps proposed in this paper can accurately reveal the cognitive levels, levels and structural characteristics of the evaluators corresponding to different comments, providing a powerful decision-making reference for improving teaching and learning and implementing precision teaching.

[0097] It also includes a system for implementing the method of using comments to construct an evaluator's cognitive structure map of the present invention, including an evaluator's comment encoding module, an encoding result visualization module, and an evaluator's cognitive structure analysis module connected in sequence.

[0098] The present invention performs structured coding of the evaluator's comments based on the evaluation criteria and knowledge system; visualizes the coding results as the evaluator's cognitive structure map through mathematical statistics and weight calculation of nodes and lines; analyzes the corresponding knowledge level, level and structural characteristics according to the evaluator's cognitive structure map, and provides a powerful decision-making reference for improving teaching and learning and implementing precision teaching. The application objects of the present invention, which are determined based on the cognitive framework theory, are: the comments and texts generated by learners who have mastered similar knowledge and skills in a certain professional field when participating in professional evaluation activities.

[0099] It will be apparent to those skilled in the art that various modifications and variations may be made to the specific embodiments of the present invention description without departing from the scope or spirit of the present invention. Other embodiments derived from the present invention description will be apparent to those skilled in the art. The present application description and examples are exemplary only.

Claims

1. A method for constructing an evaluator's cognitive structure map using comments, characterized in that: include: 1) Structural coding of the evaluator's comments based on the evaluation criteria and knowledge system; 2) Visualize the coding results into cognitive structure maps through mathematical statistics and weight calculation of nodes and lines; specifically including: 2.1) Data processing: Set each comment of each evaluator as a node unit. All the comment statements of all dimensions in a node unit are for the same person being evaluated, so these dimensions are interrelated; the dimensions between different nodes are not for the same person being evaluated, so they are not related; in a node unit, the codes of related dimensions that appear in the coding set of the dimension and appear simultaneously are all "1"; All comments of a certain evaluator are grouped into a group unit, in which there are different comments of the evaluator on different evaluated persons; the union of the dimension coding sets of all the sub-units in a group unit comprehensively reflects the knowledge level of the evaluator in the framework of professional or occupational knowledge, as well as his cognition of such knowledge; Create a matrix of the group units corresponding to each coded evaluator to quantify the correlation between the size and dimension of each code; suppose an evaluator wrote m reviews and defined the cognitive framework of a certain profession or occupation as n knowledge elements f1…f n Composition, where f represents a specific coding element, the following matrix A can be listed; where a mn It represents the coding value of the nth dimension in the mth review written by the evaluator. If its value is "1", it means that the knowledge element of the nth dimension is reflected in the mth review. If its value is "0", it means that the knowledge element of the nth dimension is not reflected in the mth review. The rows of the matrix represent the coding of the same review in each dimension, and the columns of the matrix represent the coding of a certain dimension in different reviews. The coded data of each group unit, i.e., the matrix, is displayed as a cognitive structure map in the form of a weighted network of nodes and lines: nodes represent dimensions, i.e., knowledge elements; the size of the node represents its weight, reflecting the number of times the knowledge element appears in the comments of the group unit; in the same node unit, i.e., the row of the matrix, the dimensions coded as "1" are colinear with each other, indicating that these dimensions are related in the evaluator's cognition; related dimensions are connected by lines, and the thickness of the lines represents their weight, reflecting the number of times the related dimensions appear together in the node units of a group unit; In this way, the cognitive structure map of the evaluator is constructed; each evaluator has his own unique cognitive structure map, which represents his understanding of knowledge elements and the correlation frequency of these knowledge elements, and further reflects his understanding of professional or occupational knowledge and the corresponding cognitive structure; ①Node weight calculation: D i Represents knowledge element f i The weight of D i Equal to f of matrix A i The total number of times "1" appears in the column, that is D i = a 1i + a 2i + a 3i + a 4i + … + a mi Formula 1 ② Connection weight calculation: Let f i , f j There are two knowledge elements, F i,，j Represents knowledge element f i and f j The total number of associations in a group unit, Indicates that in the kth section, the knowledge element f i and f j Whether to associate, where k = 1, 2, ..., m; 2.2) Data visualization: Draw a cognitive structure map of corresponding proportions according to the weight values ​​of nodes and lines; 3) Analyze the corresponding knowledge level, level and structural characteristics according to the evaluator's cognitive structure map to provide a powerful decision-making reference for improving teaching and learning and implementing precision teaching.

2. The method for constructing an evaluator's cognitive structure map using comments according to claim 1, characterized in that: The step 1) specifically includes: 1.1) Review collection: Collect reviews written by evaluators and remove invalid reviews; 1.2) Determine the dimensions: Construct the cognitive framework of a certain profession or occupation of the evaluator based on the knowledge and cognition that the evaluator should have, and divide the dimensions of knowledge elements as the basis for comment coding; that is, define the knowledge elements of the cognitive framework of a certain profession or occupation of the evaluator, from f1…f n Composition, where f represents a specific coded element in a specified time window, which is a specific form of some skills, knowledge, values, identity or epistemology, and n dimensions are used to represent n knowledge elements; 1.3) Comment encoding: Use binary values ​​"0" and "1" to encode the text in the comment to determine whether a comment contains the above dimensions; if the comment contains words or phrases representing the above dimensions, mark "1" on the dimension, and mark "0" on the dimension not involved, and finally convert each comment into a set of "1" and "0" codes of the dimension.

3. The method for constructing an evaluator's cognitive structure map using comments according to claim 1, characterized in that: After data visualization, cognitive structure analysis is performed: the coded data of all evaluators are visualized to obtain the cognitive structure map of each evaluator; in order to compare the differences in cognitive structures between different evaluators, their cognitive structure maps are subtracted to more clearly reflect the differences in their knowledge level and cognitive level.

4. A system for implementing the method of constructing an evaluator's cognitive structure map using comments as claimed in claim 1, characterized in that: It includes an evaluator comment coding module, a coding result visualization module, and an evaluator cognitive structure analysis module which are connected in sequence; The evaluator comment coding module performs structured coding on the evaluator's comments according to the evaluation criteria and knowledge system; specifically, it includes the comment collection submodule, dimension determination submodule, and comment coding submodule which are connected in sequence; The review collection submodule collects the reviews written by the reviewers and removes invalid reviews; The dimension determination submodule constructs the cognitive framework of a certain profession or occupation of the evaluator according to the knowledge and cognition that the evaluator should have, and divides the dimensions of knowledge elements as the basis for comment coding; that is, the knowledge elements of the cognitive framework of a certain profession or occupation to which the evaluator belongs are defined, which are composed of f1…f n Composition, where f represents a specific coded element in a specified time window, which is a specific form of some skills, knowledge, values, identity or epistemology, and n dimensions are usually used to represent n knowledge elements; The comment encoding submodule encodes the text in the comment with binary values ​​"0" and "1" to determine whether a comment contains the above dimensions; if there are words or phrases representing the above dimensions in the comment, then the dimension is marked as "1", and the dimension not involved is marked as "0", and finally each comment is converted into a set of "1" and "0" codes of the dimension; The coding result visualization module visualizes the coding result into a cognitive structure map through mathematical statistics and weight calculation of nodes and lines, including a data processing submodule and a data visualization submodule connected in sequence; The data processing submodule sets each comment of each evaluator as a section unit. The comment statements of all dimensions in a section unit are for the same person being evaluated, so these dimensions are interrelated; the dimensions between different sections are not for the same person being evaluated, so they are not related; in a section unit, the codes of related dimensions that appear in the dimension code set and appear simultaneously are all "1"; All comments of a certain evaluator are grouped into a group unit, in which there are different comments of the evaluator on different evaluated persons; the union of the dimension coding sets of all the sub-units in a group unit comprehensively reflects the knowledge level of the evaluator in the framework of professional or occupational knowledge, as well as his cognition of such knowledge; Create a matrix of the group units corresponding to each coded evaluator to quantify the correlation between the size and dimension of each code; suppose an evaluator wrote m reviews and defined the cognitive framework of a certain profession or occupation as n knowledge elements f1…f n Composition, where f represents a specific coding element, the following matrix A can be listed; where a mn It represents the coding value of the nth dimension in the mth review written by the evaluator. If its value is "1", it means that the knowledge element of the nth dimension is reflected in the mth review. If its value is "0", it means that the knowledge element of the nth dimension is not reflected in the mth review. The rows of the matrix represent the coding of the same review in each dimension, and the columns of the matrix represent the coding of a certain dimension in different reviews. The coded data of each group unit, i.e., the matrix, is displayed as a cognitive structure map in the form of a weighted network of nodes and lines: nodes represent dimensions, i.e., knowledge elements; the size of the node represents its weight, reflecting the number of times the knowledge element appears in the comments of the group unit; in the same node unit, i.e., the row of the matrix, the dimensions coded as "1" are colinear with each other, indicating that these dimensions are related in the evaluator's cognition; related dimensions are connected by lines, and the thickness of the lines represents their weight, reflecting the number of times the related dimensions appear together in the node units of a group unit; In this way, the cognitive structure map of the evaluator is constructed; each evaluator has his own unique cognitive structure map, which represents his understanding of knowledge elements and the correlation frequency of these knowledge elements, and further reflects his understanding of professional or occupational knowledge and the corresponding cognitive structure; ①Node weight calculation: D i Represents knowledge element f i The weight of D i Equal to f of matrix A i The total number of times "1" appears in the column, that is D i = a 1i + a 2i + a 3i + a 4i + … + a mi Formula 1 ② Connection weight calculation: Let f i , f j There are two knowledge elements, F i,j Represents knowledge element f i and f j The total number of associations in a group unit, Indicates that in the kth section, the knowledge element f i and f j Whether it is associated; where k = 1, 2, ..., m; The data visualization submodule draws a cognitive structure map of corresponding proportions according to the weight values ​​of nodes and lines; The evaluator's cognitive structure analysis module analyzes the corresponding knowledge level, level and structural characteristics according to the evaluator's cognitive structure map, providing a powerful decision-making reference for improving teaching and learning and implementing precise teaching.

Images