An algorithm for associating learning performance with teaching resources based on big data analysis

Abstract

The present invention relates to the field of data statistics, and provides an algorithm for associating learning performance with teaching resources based on big data analysis, comprising the following steps: obtaining student performance information and educational resource information through big data; processing missing values ​​and / or outliers of the acquired performance information and educational resource information, and establishing a linear regression model; performing correlation analysis and calculation on the linear regression model to obtain a calibrated data model; quantifying the correlation indicators in the data model, and performing graphical processing to obtain graphical data information; generating an association rule analysis model based on spark, obtaining preliminary weights of indicators that affect student performance, and generating an indicator weight adjustment module. In combination with the educational resource situation in the actual region, the indicator weights are adjusted to obtain the educational resource weights that meet the corresponding region, so as to achieve personalized guidance data, form a relatively complete and accurate data analysis, and then improve the students' performance.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to an algorithm for associating learning performance with teaching resources based on big data analysis. Background Art

[0002] Student performance is an important indicator of education quality, which directly reflects the results and level achieved by students in the learning process. The quality of student performance not only represents the individual learning ability of students, but also reflects the quality of school education. Analyzing the factors affecting students' academic performance and learning process and finding out the factors affecting students' performance can help improve students' learning effects and teaching quality.

[0003] At present, the existing analysis of factors affecting student performance is only based on traditional factors and cannot comprehensively analyze potential influencing factors; and traditional association rule mining algorithms cannot adapt well to the requirements of the rapid growth of modern data, and it is necessary to find breakthroughs on distributed and parallel computing platforms.

[0004] For example, the invention application with publication number: CN111199314A, entitled "Method for analyzing factors affecting middle school students' grades based on C4.5 algorithm", discloses a method for analyzing factors affecting middle school students' grades based on C4.5 algorithm, which is applied to the field of big data processing. It aims to address the problems that the prior art cannot objectively analyze the factors affecting students' grades, and schools and teachers cannot effectively and targetedly improve teaching methods and educational models. The method of the present invention obtains student performance data and organizes it into a data table; the columns of the data table are variables, and the rows are students; the variables include at least grade, and the grade is used as the target variable; then the C4.5 decision tree algorithm is used to analyze all variables to obtain the main characteristic variables that ultimately affect the current student's grades; it can objectively mine a large number of characteristics that affect students' grades, and plays an important supporting role in improving the educational model of schools and teachers.

[0005] For another example, the invention application with the publication number CN115577940A and the name "Method, device and equipment for acquiring teaching intervention strategies based on big data analysis" discloses the following technical solution: The invention relates to the field of data analysis, and discloses a method for acquiring teaching intervention strategies based on big data analysis, including: collecting the performance data of target students and data on factors affecting performance; performing knowledge point cognition analysis on the performance data of target students and factor feature analysis on the data on factors affecting performance, integrating the knowledge point cognition level and the factor feature analysis results to obtain the learning feature analysis results; performing a comprehensive diagnosis of the learning features on the learning feature analysis results to obtain a comprehensive learning achievement diagnosis report; identifying the students' learning difficulty factors from the comprehensive learning achievement diagnosis report, and generating teaching intervention strategies based on the learning difficulty factors. The present invention also relates to a blockchain technology, and the performance data and the data on factors affecting performance can be stored in a blockchain node. The present invention also proposes a device, equipment and medium for acquiring teaching intervention strategies based on big data analysis.

[0006] However, the aforementioned existing technologies still have certain defects, mainly: First, most of the existing technical solutions only study and analyze the impact of teaching methods, educational models and students' cognition of knowledge points on students' grades, without considering students' personal factors, school environment, school resource allocation, and potential influencing factors; Second, the existing analysis of the influencing factors of students' academic performance characteristics only outputs a characteristic diagnostic conclusion, which cannot intuitively display the size of the influencing factors and the correlation between each influencing factor, and the diagnostic report is generated based on the preset diagnostic report template and the preset report generation tool, and cannot be personalized and applied to each school. Summary of the invention

[0007] The purpose of the present invention is to provide an algorithm for associating learning performance with teaching resources based on big data analysis, aiming to improve the problem that the correlation analysis between existing educational resources and learning performance is relatively low, lacks comprehensive analysis and investigation, and leads to distorted analysis result data.

[0008] To achieve the above purpose, the present invention adopts the following technical solution: an algorithm for associating learning performance with teaching resources based on big data analysis, comprising the following steps:

[0009] Obtain students’ performance information and educational resource information through big data;

[0010] Process the missing values ​​and / or outliers of the acquired performance information and educational resource information, and establish a linear regression model;

[0011] Perform correlation analysis and calculation on the linear regression model to obtain a calibrated data model;

[0012] Quantify the correlation indicators in the data model and process them into charts to obtain chart data information;

[0013] Based on the Spark-based association rule mining and analysis algorithm, the information of educational resources is associated with student scores;

[0014] By mining association rules through the FP-Growth algorithm and interpreting the mined association rule information, we can obtain the correlation between school resources and student scores, and obtain the non-numerical influencing factors associated with student scores;

[0015] Based on the output results of the chart data information and the output results of the non-numerical influencing factors, obtain the educational resource information related to the student's performance;

[0016] Calculate the weighted scores of the category information of educational resources to obtain the correlation information between the student's score ranking and the weighted score ranking, and obtain the closeness information between the student's score and the category information of educational resources;

[0017] Based on real educational resource data and closeness information, simulated student performance information is calculated;

[0018] The simulated score information is compared with the real student score. If the simulated score information is different from the real student score, the weight value is modified and recalculated; if the simulated score information is the same as the real student score, the last calculated data of the closeness information is output.

[0019] Preferably, the student's performance information and the educational resource information include but are not limited to student personal information, student performance information, student class information, teacher information, classroom class information, school resource information and school resource allocation information.

[0020] Preferably, the missing values ​​are processed by one or more of deleting variables, filling statistics or filling dummy variables;

[0021] The outlier value is replaced by the average value or median value calculated by the obtained data, and then the replaced numerical information is normalized by a clustering algorithm to obtain normalized data information.

[0022] Preferably, data sampling is performed on the standard data information to obtain a training set and a test set of data;

[0023] A linear regression model is generated based on the training set, and then checked with the test set to obtain an approved linear regression model.

[0024] Preferably, the linear regression model is calibrated based on the improved Pearson correlation coefficient calculation and the Spearman correlation coefficient calculation to obtain a calibrated linear regression model.

[0025] Preferably, the chart data information is a wide table and / or a view table of student information, influencing indicators and student performance.

[0026] Preferably, the spark-based association rule mining and analysis algorithm comprises the following steps:

[0027] Information about educational resources is associated with student performance;

[0028] By mining association rules through the FP-Growth algorithm and interpreting the mined association rule information, we can obtain the correlation between school resources and student scores, and obtain the non-numerical influencing factors associated with student scores;

[0029] Based on the output results of the chart data information and the output results of the non-numerical influencing factors, obtain the educational resource information related to the student's performance;

[0030] The category information of educational resources is weighted and scored to obtain the correlation information between the student performance ranking and the weighted score ranking, and the closeness information between the student performance and the category information of educational resources is obtained.

[0031] Preferably, the category information of the educational resources includes but is not limited to human resources, material resources and financial resources; the human resources include but are not limited to the number of schools, class size, number of teachers, student-teacher ratio and teacher gender ratio; the material resources include but are not limited to the number of student computers, the number of interactive whiteboards, the number of school books, school area and the number of school buildings; the financial resources include but are not limited to the sources of school funding and the proportion of school activity funding.

[0032] Preferably, the weight value is modified by an indicator weight adjustment module, and the indicator weight adjustment module includes the following steps:

[0033] Constructing a judgment matrix to form a comparison matrix between the category information in the educational resources;

[0034] ;

[0035] is the element corresponding to row i and column j in matrix A; n is an integer greater than zero; based on the constructed judgment matrix, the weight algorithm management module in the indicator weight adjustment module is called to determine the relative importance of each category information, compare the category information pairwise, and assign values ​​according to the importance scale to form an importance scale matrix;

[0036] The weight value of the indicator weight adjustment module is calculated as follows:

[0037] ;

[0038] Where i is the quantity value corresponding to the number of category information, is the geometric mean of each row in the importance scaling matrix.

[0039] Preferably, a consistency check is performed on the indicator weights;

[0040] ;

[0041] ;

[0042] Among them, RI is the random consistency index associated with the total number of educational resource categories; CI is the consistency index; CR is the consistency ratio; γ is the maximum eigenvalue of the judgment matrix;

[0043] When CR<0.1, it means that the judgment matrix is ​​reasonable and the weight value of the output result is qualified.

[0044] After adopting the above technical solution, the present invention has the following advantages compared with the background technology:

[0045] 1. In terms of analysis results, the results are presented in the form of graphical graphs or tables, which makes it easier to intuitively understand the correlation between student performance and influencing factors such as school resources. It is possible to clearly understand the categories of factors that affect student performance, and to better utilize one's own resource advantages and potential resources to improve student performance.

[0046] 2. Use big data to apply association rule mining algorithms to conduct association analysis on student scores and educational resource inclinations. This can provide guiding, directive or deterministic analysis results before educational resource allocation is launched, greatly reducing the risks that may be caused by school educational resource adjustments, making the allocation of school educational resources more targeted. The purpose of reasonable resource allocation can be achieved through accurate analysis of data.

[0047] 3. When collecting data, we will collect data on factors that may affect students' grades, and then generate an association rule analysis model based on Spark to obtain the preliminary weights of indicators that affect students' grades, and generate an indicator weight adjustment module. Combined with the actual educational resources in the region, we can continuously adjust the indicator weights to obtain the educational resource weights that meet the corresponding regions, achieve personalized or targeted guidance data, form a relatively complete and accurate data analysis, and then significantly improve students' grades. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1A flowchart of an algorithm for associating learning performance with teaching resources based on big data analysis according to the present invention;

[0049] Figure 2 A dictionary tree diagram of the algorithm for associating learning performance with teaching resources based on big data analysis according to the present invention; DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0051] In addition, it should be noted that the terms "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element of the present invention must have a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0052] When an element is referred to as being “fixed to” or “disposed on” or “provided on” another element, it may be directly on the other element or indirectly on the other element. When an element is referred to as being “connected to” another element, it may be directly connected to the other element or indirectly connected to the other element.

[0053] Unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the invention can be understood according to specific circumstances. Example

[0054] Please refer to Figure 1 As shown, this implementation provides an algorithm for associating learning performance with teaching resources based on big data analysis, including the following steps:

[0055] S101. Obtain students' performance information and educational resource information through big data.

[0056] Specifically, student performance information and educational resource information include but are not limited to student personal information, student performance information, student class information, teacher information, classroom class information, school resource information, and school resource allocation information. This data can be extracted from the Ministry of Education’s government portal website to obtain regional education statistics, extract key data from text content based on the NLP algorithm (natural language processing, full name: Natural Language Processing), and complement it with the Chinese sample of PISA data (Program for International Student Assessment), and merge the two parts of data to form the basic data required above. This achieves a relatively complete acquisition of the factors that may affect student performance and ensures the comprehensiveness of the analyzed data.

[0057] S102: Process the missing values ​​and / or outliers of the acquired performance information and educational resource information, and establish a linear regression model.

[0058] Specifically, missing values ​​are processed by one or more of deleting variables, filling statistics, or filling dummy variables to avoid the impact of missing data on the results of the linear regression model. Furthermore, the data in the linear regression model is processed for outliers, that is, the outliers are replaced by the mean or median value calculated by the obtained data, and then the replaced numerical information is normalized by a clustering algorithm to obtain standardized data information. The obtained standardized data information can make the data on the same dimension, which can improve the accuracy of linear regression model recognition and avoid the impact of values ​​deviating from the normal range in the data set on the results of the linear regression model.

[0059] Data sampling is performed on the standardized data information to obtain the training set and test set of the data; a linear regression model is generated based on the training set, and then the test set is used to verify the approved linear regression model. Through data sampling, the linear regression model can be checked for recall and precision, that is, the recall and precision of the linear regression model can be verified to ensure the accuracy of the data results obtained.

[0060] S103, performing correlation analysis and calculation on the linear regression model to obtain a calibrated data model.

[0061] Specifically, the linear regression model can be calibrated based on the improved Pearson correlation coefficient calculation and the Spearman correlation coefficient calculation to obtain a calibrated linear regression model. The Pearson Correlation coefficient, also known as the Pearson Product-moment Correlation Coefficient (PPMCC\PCCs), is used to measure the linear correlation between two variables x and y, and its value is between -1 and 1. The cosine similarity mentioned above is only related to the direction of the vector, but it will be affected by the translation of the vector. In the angle cosine formula, if x is translated to x+1, the cosine value will change. The Pearson Correlation coefficient will not be affected by this situation, and sometimes the Pearson Correlation coefficient is also directly called the correlation coefficient.

[0062] The improved Pearson correlation coefficient combines the Pearson correlation coefficient and the Spearman correlation coefficient to form a supplementary and optimized correlation coefficient calculation method, which can comprehensively measure qualitative or quantitative educational resources and more accurately reflect the relationship between educational resources and student performance. Some educational resource categories are used as examples. The specific correlation matrix is ​​shown in Table 1:

[0063] Table 1 Correlation matrix

[0064]

[0065] According to the data recorded in Table 1, the corresponding relationship values ​​can be intuitively reflected so as to know which aspects of the problem need to be optimized. The above-mentioned project categories are just enumerations and are not endless.

[0066] S104: Quantify the correlation index in the data model and process it into a chart to obtain chart data information. The chart data information is a wide table and / or a view table of student information, influencing indicators and student scores.

[0067] Specifically, the features with greater influence obtained from the above correlation matrix table can be analyzed to quantify the indicators, and the cleaned data can be developed and regularized to generate a wide table of student information-influence indicators-student performance, and the relationship between the indicators can be graphically displayed.

[0068] For ease of understanding, the table is marked by way of example. The table header is part of the educational resource data, and the table data for the statistics of six students is shown in Table 2. Note that when calculating, it is necessary to perform statistical calculations on all educational resources and all students. The following table is only for example.

[0069] Table 2 Quantitative indicators

[0070]

[0071] Table 2 forms an association based on the quantitative indicator values ​​and the results of the student's performance values. The specific influencing factors are not limited to the items listed in Table 2. They can be obtained based on the actual situation in Table 1, and then a corresponding analysis is formed for each student.

[0072] The images can be presented through topographic maps, bar graphs, pie charts and other graphics, so that students can intuitively understand the specific categories of the impact of school educational resources on student performance, so as to make targeted adjustments or learning.

[0073] S105. Based on the association rule mining and analysis algorithm of Spark, different weights are added to the icon data information to obtain the closeness information between student performance and educational resources. The category information of educational resources includes but is not limited to human resources, material resources and financial resources; human resources include but are not limited to the number of schools, class size, number of teachers, student-teacher ratio and teacher gender ratio; material resources include but are not limited to the number of student computers, the number of interactive whiteboards, the number of school books, school area and the number of school buildings; financial resources include but are not limited to the source of school funds and the proportion of school activity funds.

[0074] Apache Spark is a unified analytics engine for large-scale data processing. Spark is a scheduling, monitoring, and distribution engine for big data. It is a fast and general cluster computing platform. Spark extends the popular MapReduce model. One of the main features Spark provides is the ability to run calculations in memory, but the system is also more efficient than MapReduce for complex applications that run on disk. Spark's MLlib is a machine learning component that comes in handy in big data processing. It eliminates the need to use multiple tools, one for processing and one for machine learning.

[0075] Spark provides data engineers and data scientists with a powerful, unified engine that is fast and easy to use. Spark's calculations are real-time and have less latency due to its in-memory calculations.

[0076] Specifically, the association rule mining and analysis algorithm based on Spark includes the following steps:

[0077] The information of educational resources is associated with student scores; the association rules are mined through the FP-Growth algorithm, and the mined association rule information is interpreted. The FPGrowth (Frequent Pattern Growth) algorithm is an association analysis algorithm proposed by Han Jiawei in 2000. He adopted the following divide-and-conquer strategy: the database providing frequent item sets is compressed into a frequent tree (FP tree), but the item set association information is still retained. The biggest difference between this algorithm and the Apriori algorithm is two points: it does not generate candidate sets and only needs to traverse the database twice.

[0078] The process of FP-growth algorithm is: first construct FP tree, and then use it to mine frequent itemsets. When constructing FP tree, it is necessary to scan both sides of the data set. The first scan is used to count the frequency, and the second scan is to consider the frequent itemsets.

[0079] The specific steps to implement the FP-Growth algorithm are as follows:

[0080] 1. Sequentially encode the class size, number of teachers, number of computers, etc. in the category information of educational resources into I1, I2, I3 to In (n is a natural number greater than 0) to generate an educational resource coding table.

[0081] Second, the educational resources of the students are grouped together. For example, the educational resources of the student coded T001 include I1, I2 and I5; the educational resources of the student coded T002 include I2 and I4; the educational resources of the student coded T003 include I2 and I3; the educational resources of the student coded T004 include I1, I2 and I4. In this way, all students are grouped together by student codes and educational resource codes to generate the original data table.

[0082] 3. Establish a frequent pattern tree (FP-tree) compressed data set to represent each item. For example: the number of educational resources I1 is 6; the number of educational resources I2 is 7; the number of educational resources I3 is 6; in this way, the number statistics of all educational resources are completed and a frequency format data table is generated.

[0083] Fourth, based on the generated frequency format data, sort each educational resource in the educational resource data set to establish the smallest possible dictionary tree and generate an educational resource sorting table. That is, re-sort the educational resources in the original data table, for example, sort the educational resources from high to low in frequency. After sorting, you can also intuitively know the impact of specific categories of educational resources on academic performance, which is also the sorting of influencing factors.

[0084] 5. If Figure 2As shown, obtain the conditional pattern base of the current educational resource, locate a certain node, and find the two prefixes corresponding to the node, which is the conditional pattern base of the node in the frequent pattern tree. If an educational resource does not exist on a leaf node, then the educational resource does not have a conditional pattern base, and a conditional pattern base table is generated.

[0085] 6. Obtaining frequent item sets from the conditional pattern base helps discover the associations between educational resources and enables the formulation of reasonable allocation of educational resources.

[0086] To facilitate understanding, five educational resources, namely class size, number of teachers, number of computers, funding sources and gender ratio of teachers, and seven student codes, namely T001, T002, T003, T004, T005, T006 and T007, are used as examples, and the following Table 3 Educational resource codes, Table 4 Original data, Table 5 Frequency format data, Table 6 Educational resource rankings and Table 7 Conditional model bases are generated:

[0087] Table 3 Educational resource codes

[0088]

[0089] Table 4 Raw data

[0090]

[0091] Table 5 Frequency format data

[0092]

[0093] Table 6 Ranking of educational resources

[0094]

[0095] Table 7 Conditional mode base

[0096]

[0097] In the above, the specific process for generating frequent item sets is as follows:

[0098] First, for each educational resource item in the FP-tree, find its corresponding conditional pattern base (consistent with the above steps to generate the conditional pattern base). Then build a new conditional FP-tree based on the conditional pattern base. Finally, continue to recursively mine frequent item sets on the new conditional FP-tree. The recursive process will generate new frequent item sets at each level, and finally obtain all frequent item sets and generate the frequent item set table 8 of educational resources.

[0099] Table 8 Frequent item sets of educational resources

[0100]

[0101] Furthermore, the correlation between school resources and student performance is obtained, and non-numerical influencing factors associated with student performance are acquired; based on the output results of the chart data information and the output results of the non-numerical influencing factors, educational resource information associated with student performance is obtained.

[0102] Specifically, based on the output results of the numerical transaction processing module and the output results of the non-numerical transaction processing module, the educational resources closely related to the student's performance are output. According to the influencing factors (i.e. the ranking of educational resources), different educational resources are processed with different weights to obtain the weighted scores of educational resources of each school.

[0103] It has been realized to calculate the weighted scores of the category information of educational resources, obtain the correlation information between the student performance ranking and the weighted score ranking, and obtain the closeness information between the student performance and the category information of educational resources.

[0104] Specifically, the weighted coefficient is calculated as follows: weighted coefficient = training set educational resource ratio ÷ actual ratio; based on the above example, the following weighted coefficient table in Table 9 can be obtained:

[0105] Table 9 Weighting coefficients

[0106]

[0107] S106. Calculate and obtain simulated student performance information based on real educational resource data and closeness information; compare the simulated performance information with the real student performance. If the simulated performance information is different from the real student performance, modify the weight value and recalculate; if the simulated performance information is the same as the real student performance, output the last calculated data of the closeness information.

[0108] The weight value is modified by an indicator weight adjustment module, which includes the following steps:

[0109] Constructing a judgment matrix to form a comparison matrix between the category information in the educational resources;

[0110] ;

[0111] is the element corresponding to row i and column j in matrix A, and n is an integer greater than zero; based on the constructed judgment matrix, the weight algorithm management module in the indicator weight adjustment module is called to determine the relative importance of each category information, compare the category information pairwise, and assign values ​​according to the importance scale to form an importance scale matrix;

[0112] The weight value of the indicator weight adjustment module is calculated as follows:

[0113] ;

[0114] Where i is the quantity value corresponding to the number of category information, is the geometric mean of each row in the importance scaling matrix.

[0115] Conducting consistency checks on the indicator weights;

[0116] ;

[0117] ;

[0118] Among them, RI is the random consistency index associated with the total number of educational resource categories; CI is the consistency index; CR is the consistency ratio; γ is the maximum eigenvalue of the judgment matrix;

[0119] When CR<0.1, it means that the judgment matrix is ​​reasonable and the weight value of the output result is qualified, that is, the output result can be used as the weight of educational resources.

[0120] Specifically, the algorithm for associating learning results with teaching resources based on big data analysis is calculated based on the association rule mining analysis algorithm of Spark, which has the effect of faster processing speed and higher ease of use; and the application of image processing methods can more intuitively show the relationship between various factors and student performance. When applying the association mining algorithm, a weight adjustment module is generated, and when it is finally applied, it is continuously adjusted based on actual data, so that the model results are closer to the real data, ensuring the personalization and accuracy of the model in different regions.

[0121] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An algorithm for associating learning performance with teaching resources based on big data analysis, characterized in that: The following steps are involved: Obtain students’ performance information and educational resource information through big data; Process the missing values ​​and / or outliers of the acquired performance information and educational resource information, and establish a linear regression model; Perform correlation analysis and calculation on the linear regression model to obtain a calibrated data model; Quantify the correlation indicators in the data model and process them into charts to obtain chart data information; Based on the Spark-based association rule mining and analysis algorithm, the information of educational resources is associated with student scores; By mining association rules through the FP-Growth algorithm and interpreting the mined association rule information, we can obtain the correlation between school resources and student scores, and obtain the non-numerical influencing factors associated with student scores; Based on the output results of the chart data information and the output results of the non-numerical influencing factors, obtain the educational resource information related to the student's performance; Calculate the weighted scores of the category information of educational resources to obtain the correlation information between the student score ranking and the weighted score ranking, and obtain the closeness information between the student score and the category information of educational resources; Based on real educational resource data and closeness information, simulated student performance information is calculated; Compare the simulated score information with the real student score. If the simulated score information is different from the real student score, modify the weight value and recalculate; if the simulated score information is the same as the real student score, output the last calculated data of the closeness information; The missing values ​​are processed by one or more of deleting variables, filling statistics or filling dummy variables; The outlier value is replaced by the average value or median value calculated by the obtained data, and then the replaced numerical information is normalized by a clustering algorithm to obtain normalized data information; Performing data sampling on the standard data information to obtain a training set and a test set of data; Generate a linear regression model based on the training set, and then check it with the test set to obtain an approved linear regression model; The linear regression model is calibrated based on the improved Pearson correlation coefficient calculation and the Spearman correlation coefficient calculation to obtain a calibrated linear regression model.

2. The algorithm for associating learning performance with teaching resources based on big data analysis according to claim 1 is characterized by: The student's performance information and the educational resource information include student personal information, student performance information, student class information, teacher information, classroom class information, school resource information and school resource allocation information.

3. The algorithm for associating learning performance with teaching resources based on big data analysis according to claim 1 is characterized in that: The chart data information is a wide table and / or a view table of student information, influencing indicators and student performance.

4. The algorithm for associating learning performance with teaching resources based on big data analysis according to claim 1 is characterized in that: The category information of the educational resources includes human resources, material resources and financial resources; the human resources include the number of schools, class size, number of teachers, student-teacher ratio and teacher gender ratio; the material resources include the number of student computers, the number of interactive whiteboards, the number of school books, school area and the number of school buildings; the financial resources include the source of school funds and the proportion of school activity funds.

5. The algorithm for associating learning performance with teaching resources based on big data analysis according to claim 1 is characterized in that: The weight value is modified by an indicator weight adjustment module, and the indicator weight adjustment module includes the following steps: Constructing a judgment matrix to form a comparison matrix between the category information in the educational resources; ； is the element corresponding to row i and column j in matrix A, and n is an integer greater than zero; based on the constructed judgment matrix, the weight algorithm management module in the indicator weight adjustment module is called to determine the relative importance of each category information, compare the category information pairwise, and assign values ​​according to the importance scale to form an importance scale matrix; The weight value of the indicator weight adjustment module is calculated as follows: ； Where i is the quantity value corresponding to the number of category information, is the geometric mean of each row in the importance scaling matrix.

6. The algorithm for associating learning performance with teaching resources based on big data analysis according to claim 5 is characterized in that: Performing consistency check on the weight values ​​of the indicator weight adjustment module; ； ； Among them, RI is the random consistency index associated with the total number of educational resource categories; CI is the consistency index; CR is the consistency ratio; γ is the maximum eigenvalue of the judgment matrix; When CR<0.1, it means that the judgment matrix is ​​reasonable and the weight value of the output result is qualified.

Images