Student score online monitoring method and system based on improved FA algorithm optimized SVR

Abstract

The invention discloses a student score online monitoring method and system based on improved FA algorithm optimization SVR, and relates to the technical field of student score monitoring, and the method comprises the following steps: 1, carrying out the normalization processing of initial sample data; 2, setting a support vector regression model and improved firefly algorithm initial parameters; 3, initializing a population; 4, decoding operation is executed, and the fitness is calculated; 5, updating the positions of the fireflies; 6, updating the optimal solution; and 7, adding one to the number of iterations, and returning to the step 4. According to the method, model parameters are optimized by introducing the improved firefly algorithm, the prediction accuracy of the model is improved, compared with a traditional method, the improved firefly algorithm simulates glitters and mutual attraction behaviors of fireflies, better solutions are continuously searched in a parameter space, finally, a parameter combination enabling the model prediction error to be minimum is found, and the prediction accuracy of the model is improved. Therefore, the prediction accuracy of the SVR model is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of student performance monitoring, and in particular to an online student performance monitoring method and system based on an improved FA algorithm to optimize SVR. Background Art

[0002] Student performance monitoring is a method used to study students' learning status. FA algorithm, namely firefly algorithm, and SVR, namely vector regression, can be used to predict students' performance. Traditional statistical models or simple data-driven models, such as firefly algorithm, often have problems such as high computational complexity and insufficient prediction accuracy when processing high-dimensional feature data. Especially when considering that students' performance is affected by multiple learning behavior characteristics, it is difficult to effectively capture these complex relationships, thus affecting the accuracy and real-time performance prediction.

[0003] The defects of existing methods of monitoring student performance are:

[0004] 1. The prior art US07311524B2 discloses a system and method for evaluating student performance. This technology does not have the function of continuously searching for better solutions in the parameter space, which easily leads to inaccurate student evaluation results. Therefore, a method for online monitoring of student performance based on an improved FA algorithm to optimize SVR that can continuously optimize parameters and continuously search for better solutions is needed to solve this problem.

[0005] 2. The prior art CN115757648A discloses a multi-modal grade chain-uploading method and system based on the alliance chain DID. This technology cannot collect students' basic information and learning behavior characteristics in real time through multiple channels such as the school's information management system, online learning platform, and classroom interactive system, which affects the accuracy and real-time performance of grade prediction. Therefore, a student grade online monitoring method based on an improved FA algorithm to optimize SVR that can obtain student grade information through multiple channels in real time online is needed to solve this problem.

[0006] 3. Prior art CN112150327A discloses a test system for student physical fitness testing. This technology does not have a mechanism for updating the radius of the best solution. It is easy to ignore some factors that affect student performance when calculating student performance, resulting in predictions of student performance under local restrictive conditions rather than global objective conditions, which can easily lead to inaccurate predicted performance. Therefore, a method for online monitoring of student performance based on an improved FA algorithm to optimize SVR is needed, which can update the radius of the best solution, dynamically adjust the step size model, and use normal distribution to generate new solutions to solve this problem. Summary of the invention

[0007] One purpose of the present application is to provide a method and system for online monitoring of student performance based on an improved FA algorithm to optimize SVR, which can solve the technical problems raised in the prior art.

[0008] To achieve the above object, the present invention provides the following technical solution: an online monitoring method for student performance based on an improved FA algorithm to optimize SVR, comprising the following steps:

[0009] Step 1: Normalize the initial sample data;

[0010] Step 2: Set the initial parameters of the support vector regression model and the improved firefly algorithm;

[0011] Step 3: Initialize the population;

[0012] Step 4: Perform decoding operation and calculate fitness;

[0013] Step 5: Update the firefly position;

[0014] Step 6: Update the optimal solution;

[0015] Step 7: Increase the number of iterations by one and return to step 4;

[0016] The student performance online monitoring system based on the improved FA algorithm to optimize SVR executes steps one to seven.

[0017] Preferably, the step 1: normalizing the initial sample data, normalizing the initial sample data according to formula (1),

[0018]

[0019] where a ij is the sample data, a imin with a imax are the minimum and maximum values ​​of the learning behavior characteristics of the i-th student.

[0020] Preferably, the step 2 is: setting the initial parameters of the support vector regression model and the improved firefly algorithm, including the search range C∈[10 -3 ,10 3 ], the search range of kernel function bandwidth γ is γ∈[10 -6 ,10 1 ], the search range of the allowable error ε∈[10 -3 ,10 -1 ], initial population size N = 100, light absorption coefficient ρ = 0.2, attraction strength β = 0.8, step factor α = 0.1, current iteration number t = 1, maximum iteration number t max .

[0021] Preferably, the step three: initialize the population, encode the model parameters (C, γ, ε), each solution is represented as a firefly individual, and the i-th solution is represented as

[0022]

[0023] Preferably, the step 4: performing decoding operation, calculating fitness, and improving the firefly algorithm, (a) updating the radius based on the best solution:

[0024] Let X i is the current solution, X Gbest is the current optimal solution, then based on the radius r of the optimal solution ibest The calculation formula is:

[0025]

[0026] (b) Update step size calculation:

[0027] First method: ΔX 1ij =X Gbest -X worst ;

[0028] Second method: ΔX 2ij =(X j -X i +X r1 -X r2 );

[0029] The first update step size ΔX 1ij Used to narrow the search range. When the current solution is close to the global optimum, it helps to refine the search. The second update step size model ΔX 2ij , where X r1 and X r2 are the positions of two randomly selected fireflies, which are used to expand the search range and avoid the algorithm from falling into the local optimum. The step size used depends on the fitness ratio FR of the i-th firefly and the current best firefly. i and the fitness ratio of the entire population to the current best firefly

[0030]

[0031] FT i is the fitness function value of the i-th firefly, FT Gbest is the fitness function value of the best firefly in the current population, is the average value of the fitness function of all fireflies in the entire population, ΔX ij The choices are as follows:

[0032]

[0033] (c) Generate new solutions using normal distribution:

[0034] The new solution generation formula is:

[0035]

[0036] Indicates that by updating the step size ΔX ij and the position of the new solution generated by the random number generator randn., randn. represents a random number generator with a normal distribution.

[0037] Preferably, the step 5: updating the firefly position, its attraction β ibest The calculation formula is:

[0038]

[0039] FireflyX i To X Gbest Move, updated position:

[0040]

[0041] Indicates that by updating the step size ΔX iGbest and the position of the new solution generated by the random number generator randn., randn. represents the random number generator of the normal distribution.

[0042] Preferably, the step six: update the optimal solution, set t=t+1, and determine t <t max Is it true? If so, execute step 7; otherwise, end the search and output the optimal solution.

[0043] Preferably, the student performance online monitoring system based on the improved FA algorithm to optimize SVR includes a data transmission module, a parameter setting module, a model training module, an intelligent decision-making module and an abnormal alarm module, wherein the data transmission module is used to execute step 1, collect the basic information and learning behavior data of the students, and pre-process the collected data;

[0044] The parameter setting module is used to execute step 2, and set the parameters required for the monitoring module, including penalty parameter, kernel function bandwidth, allowable error, initial population size, light absorption coefficient, attraction intensity, and step factor.

[0045] Preferably, the model training module is used to execute steps 3 to 7, importing learning sample data (X train ,Y train), apply the learning sample data to train the support vector machine regression model, first initialize the model Y train =SVR(X train ), and then use the improved firefly algorithm to optimize the model parameters and find the optimal model parameters;

[0046] The intelligent decision-making module is used to import online monitoring sample data X test , use the trained model to test the performance of online monitoring samples, Y test =SVR(X test ), output the performance monitoring result Y test .

[0047] Preferably, the abnormal alarm module makes a quick judgment on the student performance monitoring results, notifies the relevant teachers of the students with abnormal performance as soon as possible, and records, views and tracks the abnormal students.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. The present invention optimizes the model parameters by introducing an improved firefly algorithm to improve the prediction accuracy of the model. Compared with the traditional method, the improved firefly algorithm simulates the flashing and mutual attraction behavior of fireflies, continuously searches for better solutions in the parameter space, and finally finds the parameter combination that minimizes the model prediction error, thereby improving the prediction accuracy of the SVR model.

[0050] 2. The present invention collects students' basic information and learning behavior characteristics in real time through multiple channels such as the school's information management system, online learning platform, and classroom interactive system. These data constitute a high-dimensional feature space, providing a rich information basis for subsequent prediction models.

[0051] 3. The present invention improves the firefly algorithm by introducing strategies such as a radius update mechanism based on the optimal solution, a dynamically adjusted step size model, and the use of a normal distribution to generate new solutions. This significantly enhances the global search capability of the firefly algorithm, effectively avoids the local optimal problem, and improves the search stability and convergence speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of the online monitoring method of student performance based on improved firefly algorithm-support vector regression of the present invention;

[0053] Figure 2 The improved firefly algorithm flow chart of the present invention;

[0054] Figure 3 The schematic diagram of the structure of the online monitoring system of student performance based on improved firefly algorithm-support vector regression of the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] See also Figure 1 and Figure 2 , an embodiment provided by the present invention: an online monitoring method for student performance based on optimizing SVR by improved FA algorithm;

[0057] The steps include:

[0058] Step 1: Normalize the initial sample data;

[0059] Step 2: Set the initial parameters of the support vector regression model and the improved firefly algorithm;

[0060] Step 3: Initialize the population;

[0061] Step 4: Perform decoding operation and calculate fitness;

[0062] Step 5: Update the firefly position;

[0063] Step 6: Update the optimal solution;

[0064] Step 7: Increase the number of iterations by one and return to step 4;

[0065] The student performance online monitoring system based on the improved FA algorithm to optimize SVR executes steps one to seven.

[0066] Step 1: Normalize the initial sample data. Normalize the initial sample data according to formula (1).

[0067]

[0068] where a ij is the sample data, a imin with a imax are the minimum and maximum values ​​of the learning behavior characteristics of the i-th student.

[0069] Step 2: Set the initial parameters of the support vector regression model and the improved firefly algorithm, including the search range C∈[10 -3 ,10 3 ], the search range of kernel function bandwidth γ is γ∈

[0070] [10 -6 ,10 1 ], the search range of the allowable error ε∈[10 -3 ,10 -1 ], initial population size N = 100, light absorption coefficient ρ = 0.2, attraction strength β = 0.8, step factor α = 0.1, current iteration number t = 1, maximum iteration number t max .

[0071] Step 3: Initialize the population and encode the model parameters (C, γ, ε). Each solution is represented by a firefly individual, and the i-th solution is represented by

[0072]

[0073] Step 4: Perform decoding operation, calculate fitness, and improve the firefly algorithm. (a) Radius update based on the best solution:

[0074] Let X i is the current solution, X Gbest is the current optimal solution, then based on the radius r of the optimal solution ibest The calculation formula is:

[0075]

[0076] (b) Update step size calculation:

[0077] First method: ΔX 1ij =X Gbest -X worst ;

[0078] Second method: ΔX 2ij =(X j -X i +X r1 -X r2 );

[0079] The first update step size ΔX 1ij Used to narrow the search range. When the current solution is close to the global optimum, it helps to refine the search. The second update step size model ΔX 2ij , where X r1 and X r2 are the positions of two randomly selected fireflies, which are used to expand the search range and avoid the algorithm from falling into the local optimum. The step size used depends on the fitness ratio FR of the i-th firefly and the current best firefly. i and the fitness ratio of the entire population to the current best firefly

[0080]

[0081] FT i is the fitness function value of the i-th firefly, FT Gbest is the fitness function value of the best firefly in the current population, is the average value of the fitness function of all fireflies in the entire population, ΔX ij The choices are as follows:

[0082]

[0083] (c) Generate new solutions using normal distribution:

[0084] The new solution generation formula is:

[0085]

[0086] Indicates that by updating the step size ΔX ij and the position of the new solution generated by the random number generator randn., randn. represents a random number generator with a normal distribution.

[0087] Step 5: Update the firefly position and its attraction β ibest The calculation formula is:

[0088]

[0089] FireflyX i To X Gbest Move, updated position:

[0090]

[0091] Indicates that by updating the step size ΔX iGbest and the position of the new solution generated by the random number generator randn., randn. represents the random number generator of the normal distribution.

[0092] Step 6: Update the optimal solution, set t = t + 1, and determine t <t max Is it true? If so, execute step 7; otherwise, end the search and output the optimal solution.

[0093] See also Figure 3 , an embodiment provided by the present invention: an online monitoring system for student performance based on the improved FA algorithm to optimize SVR;

[0094] The student performance online monitoring system based on the improved FA algorithm to optimize SVR includes a data transmission module, a parameter setting module, a model training module, an intelligent decision-making module and an abnormal alarm module. The data transmission module is used to execute step one, collect students' basic information and learning behavior data, and pre-process the collected data;

[0095] The parameter setting module is used to execute step 2 and set the parameters required for the monitoring module, including penalty parameter, kernel function bandwidth, allowable error, initial population size, light absorption coefficient, attraction intensity, and step factor.

[0096] The model training module is used to execute steps 3 to 7 and import learning sample data (X train ,Y train ), apply the learning sample data to train the support vector machine regression model, first initialize the model Y train =SVR(X train ), and then use the improved firefly algorithm to optimize the model parameters and find the optimal model parameters;

[0097] Intelligent decision-making module is used to import online monitoring sample data X test , use the trained model to test the performance of online monitoring samples, Y test =SVR(X test ), output the performance monitoring result Y test .

[0098] The abnormal alarm module makes a quick judgment on the student performance monitoring results, notifies the relevant teachers of students with abnormal performance as soon as possible, and records, views and tracks the abnormal students.

[0099] For those skilled in the art, the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for online monitoring of student performance based on an improved FA algorithm to optimize SVR, characterized in that: The steps include: Step 1: Normalize the initial sample data; Step 2: Set the initial parameters of the support vector regression model and the improved firefly algorithm; Step 3: Initialize the population; Step 4: Perform decoding operation and calculate fitness; Step 5: Update the firefly position; Step 6: Update the optimal solution; Step 7: Increase the number of iterations by one and return to step 4; The student performance online monitoring system based on the improved FA algorithm to optimize SVR executes steps one to seven.

2. The method for online monitoring of student performance based on the improved FA algorithm to optimize SVR according to claim 1 is characterized by: Step 1: normalize the initial sample data according to formula (1). where a ij is the sample data, a imin with a imax are the minimum and maximum values ​​of the learning behavior characteristics of the i-th student.

3. The method for online monitoring of student performance based on the improved FA algorithm to optimize SVR according to claim 1 is characterized by: Step 2: Set the initial parameters of the support vector regression model and the improved firefly algorithm, including the search range C∈[10 -3 ,10 3 ], the search range of kernel function bandwidth γ is γ∈[10 -6 ,10 1 ], the search range of the allowable error ε∈[10 -3 ,10 -1 ], initial population size N = 100, light absorption coefficient ρ = 0.2, attraction strength β = 0.8, step factor α = 0.1, current iteration number t = 1, maximum iteration number t max .

4. The method for online monitoring of student performance based on the improved FA algorithm to optimize SVR according to claim 3 is characterized by: Step 3: Initialize the population, encode the model parameters (C, γ, ε0, each solution is represented by a firefly individual, and the i-th solution is represented by 5. The method for online monitoring of student performance based on the improved FA algorithm to optimize SVR according to claim 3 is characterized by: Step 4: Execute decoding operation, calculate fitness, and improve the firefly algorithm. (a) Update the radius based on the best solution: Let X i is the current solution, X Gbest is the current optimal solution, then based on the radius r of the optimal solution ibest The calculation formula is: (b) Update step size calculation: First method: ΔX 1ij =X Gbest -X worst ; Second method: ΔX 2ij =(X j -X i +X r1 -X r2 ); The first update step size ΔX 1ij Used to narrow the search range. When the current solution is close to the global optimum, it helps to refine the search. The second update step size model ΔX 2ij , where X r1 and X r2 are the positions of two randomly selected fireflies, which are used to expand the search range and avoid the algorithm from falling into the local optimum. The step size used depends on the fitness ratio FR of the i-th firefly and the current best firefly. i and the fitness ratio FR of the entire population and the current best firefly Npop , FT i is the fitness function value of the i-th firefly, FT Gbest is the fitness function value of the best firefly in the current population, FR Npop is the average value of the fitness function of all fireflies in the entire population, ΔX ij The choices are as follows: (c) Generate new solutions using normal distribution: The new solution generation formula is: Indicates that by updating the step size ΔX ij and the position of the new solution generated by the random number generator randn., where randn. represents a random number generator with a normal distribution.

6. The method for online monitoring of student performance based on the improved FA algorithm to optimize SVR according to claim 5, characterized in that: Step 5: Update the firefly position, its attraction β ibest The calculation formula is: FireflyX i To X Gbest Move, updated position: Indicates that by updating the step size ΔX iGbest and the position of the new solution generated by the random number generator randn., randn. represents the random number generator of the normal distribution.

7. The method for online monitoring of student performance based on the improved FA algorithm to optimize SVR according to claim 3 is characterized by: Step 6: Update the optimal solution, set t=t+1, and determine t <t max Is it true? If so, execute step 7; otherwise, end the search and output the optimal solution.

8. An online student performance monitoring system based on an improved FA algorithm to optimize SVR, characterized by: The student performance online monitoring system based on the improved FA algorithm to optimize SVR includes a data transmission module, a parameter setting module, a model training module, an intelligent decision-making module and an abnormal alarm module. The data transmission module is used to execute step 1, collect the basic information and learning behavior data of the students, and pre-process the collected data; The parameter setting module is used to execute step 2, and set the parameters required for the monitoring module, including penalty parameter, kernel function bandwidth, allowable error, initial population size, light absorption coefficient, attraction intensity, and step factor.

9. The student performance online monitoring system based on the improved FA algorithm to optimize SVR according to claim 8 is characterized by: The model training module is used to execute steps 3 to 7, import learning sample data (X train ,Y train ), apply the learning sample data to train the support vector machine regression model, first initialize the model Y train =SVR(X train ), and then use the improved firefly algorithm to optimize the model parameters and find the optimal model parameters; The intelligent decision-making module is used to import online monitoring sample data X test , use the trained model to test the performance of online monitoring samples, Y test =SVR(X test ), output the performance monitoring result Y test .

10. The student performance online monitoring system based on the improved FA algorithm to optimize SVR according to claim 8, characterized in that: The abnormal alarm module makes a quick judgment on the student performance monitoring results, notifies the relevant teachers of students with abnormal performance as soon as possible, and records, views and tracks the abnormal students.

Images