Campus potential infectious disease intelligent prediction machine learning integrated optimization method and system
By adopting machine learning methods with multi-model integration and particle swarm optimization algorithms on campus, the problem of traditional manual judgments being difficult to accurately identify potential infectious diseases is solved, and more efficient and accurate disease prediction and prevention and control are achieved.
Patent Information
- Application Number
- CN202411731177.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-29
- Publication Date
- 2025-04-18
AI Technical Summary
It is difficult for the prior art to quickly, efficiently and accurately identify potential infectious diseases in the prevention and control of diseases on campus. Especially when faced with cases with complex symptoms or insignificant early symptoms, traditional manual judgments may cause misjudgment, leading to the spread of the disease.
An integrated optimization method for intelligent prediction of potential infectious diseases on campus is proposed. By obtaining student health status data, extracting physical characteristics, inputting multiple machine learning models (such as Bayesian model, decision tree model, support vector machine model, etc.) for prediction, and weighted calculation and particle swarm optimization algorithm are used to optimize model weights to improve prediction accuracy.
It significantly improves the accuracy and robustness of disease prediction, can better adapt to complex and heterogeneous student health data, achieve efficient identification of multiple infectious diseases, reduce the risk of disease spread, and provide intelligent and automated prediction support.
Smart Images
Figure CN119851970A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of machine learning, and in particular relates to a method and system for intelligent prediction of potential infectious diseases on campus and machine learning integrated optimization. Background Art
[0002] As a highly densely populated place, close contact and frequent interactions between students on campus provide convenient conditions for the spread of diseases, especially when facing highly contagious diseases such as respiratory diseases and skin diseases, the epidemic spreads faster and affects a wider range. Early detection and control of infectious diseases have become a key issue in ensuring campus safety and health. In order to effectively prevent and control infectious diseases, identify students' illness status in a timely manner, and make corresponding isolation or treatment recommendations, schools need a fast, efficient, and accurate disease prediction mechanism to reduce the risk of disease transmission.
[0003] At present, disease prevention and control on campus mostly rely on traditional manual reporting and health monitoring methods. Teachers and school doctors judge whether there are potential diseases based on students' health performance or through preliminary screening methods such as temperature detection. Although these methods have certain operability, they are inefficient when processing large amounts of data and complex information, and it is difficult to accurately identify the types of potential infectious diseases. Especially in the face of some cases where early symptoms are not obvious or the symptoms are complex, traditional manual judgment may lead to misjudgment, resulting in further spread of infectious diseases. Therefore, how to use modern intelligent technology to quickly analyze and predict students' health data is an urgent problem to be solved.
[0004] Machine learning technology has been widely used in health management, disease diagnosis and other fields due to its powerful data processing capabilities. By learning and modeling a large amount of health data, machine learning can realize the automatic identification and prediction of complex diseases, thus providing new solutions for campus infectious disease prevention and control. However, students' health status is complex and diverse, and the characteristics involved include not only physical symptoms (such as fever, cough, sore throat, headache, etc.), but also external influencing factors (such as mental illness, surgical recovery period, etc.). Using only a single machine learning model often cannot fully explore the intrinsic connection between these features, resulting in low prediction accuracy. Therefore, in order to better cope with this multi-variable, multi-feature disease prediction task, more advanced model integration technology is needed. Summary of the invention
[0005] In order to solve the above technical problems, the present invention proposes a method and system for intelligent prediction of potential infectious diseases on campus, which can improve the accuracy of disease prediction.
[0006] The present invention provides a method for intelligent prediction of potential infectious diseases on campus and machine learning integrated optimization, comprising:
[0007] Obtain student health status data;
[0008] Extracting a number of body features based on the health status data to obtain a feature vector;
[0009] Inputting the feature vector into different types of machine learning models to obtain corresponding prediction results;
[0010] The corresponding prediction results are weighted to obtain the optimal prediction result.
[0011] Optionally, inputting the feature vector into a preset number of machine learning models to obtain a prediction result includes:
[0012] The feature vectors are respectively input into the Bayesian model, decision tree model, support vector machine model, BP neural network model, logistic regression model and random forest model to obtain the prediction results of a single model.
[0013] Optionally, the method for obtaining the prediction result using the Bayesian model is:
[0014]
[0015] Among them, C k is the kth disease category, X is the feature vector, P(X|C k ) is the likelihood of the feature under a specific disease, P(C k ) is the prior probability of the disease, and P(X) is the total probability distribution of the feature.
[0016] Optionally, the method for obtaining the prediction result using the decision tree model is:
[0017]
[0018] Among them, Gain(X, A) is the information gain brought by feature X in data set A, HH(X) represents the entropy of the data set, A is the partition feature, and X v is the subset after partitioning.
[0019] Optionally, the method for obtaining the prediction result using the logistic regression model is:
[0020]
[0021] Among them, w is the weight vector of the model, X is the input feature, and w T is the full-time weight vector, and y is the predicted target category.
[0022] Optionally, performing weighted calculation on the prediction results to obtain the optimal prediction result includes:
[0023] Performing weighted integration on the prediction results of the individual models to obtain a comprehensive prediction result;
[0024] The particle swarm optimization algorithm is used to optimize the weight of the comprehensive prediction result to obtain the best prediction result.
[0025] Optionally, a method for optimizing the weight of the comprehensive prediction result using a particle swarm optimization algorithm is:
[0026] v i+1 =wv i +c1r1(p i，best -w i )+c2r2(g best -w i )
[0027] w i+1 =w i +v i+1
[0028] Among them, w i is the weight vector of the ith particle, v i is the velocity of the ith particle, p i,best is the particle’s own historical optimal position, g best is the global optimal position, c1, c2 are learning factors, r1, r2 are random numbers, v i+1 is the velocity of the i+1th particle, w i+1 is the weight vector of the i+1th particle.
[0029] The present invention also provides a campus potential infectious disease intelligent prediction machine learning integrated optimization system, including: a feature extraction module, a model prediction module, a model integration module and a weight optimization module;
[0030] The feature extraction module is used to obtain the student's health status data, extract a number of body features based on the health status data, and obtain a feature vector;
[0031] The model prediction module is used to input the feature vector into different types of machine learning models to obtain corresponding prediction results;
[0032] The model integration module is used to perform weighted calculation on the prediction results to obtain integrated prediction results;
[0033] The weight optimization module is used to optimize the weight of the model integration.
[0034] Compared with the prior art, the present invention has the following advantages and technical effects:
[0035] 1. Through the weighted combination of multiple models, the present invention significantly improves the accuracy and robustness of the prediction results, and can better adapt to complex and heterogeneous student health data, thereby achieving efficient identification of multiple infectious diseases.
[0036] 2. The weighting strategy of the present invention enables the method to dynamically adjust the contribution of each model according to actual needs when dealing with diseases with different characteristics, thereby improving the overall prediction effect, especially giving greater weight to some models that perform better in specific tasks, thereby enhancing the model's prediction ability for a variety of infectious diseases.
[0037] 3. The PSO algorithm of the present invention has a strong global search capability and can quickly find the optimal weight combination, so that the prediction performance of the model integration reaches the best state. Compared with traditional optimization algorithms, particle swarm optimization converges faster on complex problems and produces better results, thereby improving the overall efficiency and accuracy of disease prediction.
[0038] 4. Through more comprehensive feature extraction, the present invention can capture more potential disease characteristics, especially when the early symptoms of infectious diseases are not obvious or complex, the predictive ability of the method is significantly improved, ensuring timely and accurate identification of potential infectious diseases.
[0039] 5. The present invention improves the ability to predict infectious diseases among students by combining the specific health characteristics of the campus population, making the method more operable and practical in practical applications. At the same time, this targeted design ensures that the method can quickly make effective predictions, providing accurate decision-making support for school health management and reducing the risk of disease spread.
[0040] 6. Through automated processing, the method not only reduces the need for manual intervention and improves work efficiency, but also ensures efficient prediction under large-scale data, which is suitable for health monitoring and infectious disease early warning of thousands or even tens of thousands of students in a campus environment. At the same time, the intelligent design enables the prediction results to be generated in real time, which is conducive to the early detection and timely intervention of diseases.
[0041] 7. This scalability makes the present invention not only applicable to the prediction of campus infectious diseases, but also applicable to other public health scenarios, and has a wide range of application value. The versatility of the method ensures that it can be customized according to different scenarios and needs, thus playing a role in various disease monitoring and prevention work. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0043] Figure 1 This is a flow chart of a method for intelligent prediction of potential infectious diseases on campus and machine learning integrated optimization according to an embodiment of the present invention;
[0044] Figure 2 It is a framework diagram of a method for intelligent prediction of potential infectious diseases on campus by machine learning integration optimization according to an embodiment of the present invention. DETAILED DESCRIPTION
[0045] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0046] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0047] Ensemble Learning is a technology that combines the prediction results of multiple single models to improve the overall prediction performance. By integrating different types of machine learning models, we can make full use of the advantages of each model, make up for the shortcomings of a single model, and improve the robustness and accuracy of the prediction. Common ensemble strategies include weighted averaging, voting, stacking, etc. In particular, the weighted ensemble method can assign different weights to each model according to its performance on different tasks, thereby improving the accuracy and stability of the prediction.
[0048] In the present invention, the integrated Bayesian model, decision tree model, support vector machine model, BP neural network model, logistic regression model and random forest model are considered. By weighting the prediction results of these models, the relationship between students' health characteristics and potential infectious diseases can be captured more comprehensively. However, how to determine the optimal weight of each model is the key to model integration, and the traditional manual weight setting method is often difficult to achieve the optimal effect. Therefore, the present invention introduces a particle swarm algorithm (PSO) to optimize the model weights.
[0049] Particle swarm optimization (PSO) is a global optimization algorithm based on swarm intelligence, which simulates the behavior of bird flocks foraging, and finds the optimal solution in the search space of the problem through information sharing and self-adjustment between individuals. In the integration of machine learning models, PSO can adjust the weights of the model by continuous iteration to optimize the overall prediction performance of the integrated model. Compared with the traditional gradient descent method, PSO has the advantages of simple calculation and strong global search ability, and is particularly suitable for high-dimensional complex optimization problems. Therefore, in the present invention, PSO is used to optimize the weights of each machine learning model to further improve the accuracy of disease prediction.
[0050] The present invention proposes a machine learning integrated optimization method for intelligent prediction of potential infectious diseases on campus. Figure 1-2 As shown, the specific steps include:
[0051] Obtain student health status data;
[0052] Extracting several body features based on health status data to obtain feature vectors;
[0053] Input feature vectors into different types of machine learning models to obtain corresponding prediction results;
[0054] Perform weighted calculation on the corresponding prediction results to obtain the optimal prediction result.
[0055] The present invention proposes a machine learning integrated optimization method for intelligent prediction of potential infectious diseases on campus. By integrating Bayesian model, decision tree model, support vector machine model, BP neural network model, logistic regression model and random forest model, and optimizing model weights using particle swarm algorithm, the prediction problem caused by complex health status of students and diverse disease types is effectively solved. This method not only improves the accuracy of infectious disease prediction, but also provides intelligent and automated technical means for the prevention and control of infectious diseases on campus, which is helpful for ensuring campus safety and maintaining student health. Through the application of the invention method, potential campus infectious diseases can be identified earlier and more accurately, and a scientific basis can be provided for the formulation of relevant prevention and control measures, greatly improving the management efficiency of campus public health.
[0056] Specifically, the task of feature extraction is to extract valuable feature information from the students’ health data to form a feature vector. The specific features include the students’ fever, symptoms, psychological state, etc. Suppose the input student data is D = {d1, d2, …, d n}, each data contains 34 features, which can be expressed as:
[0057] X=(x1,x2,…,x 34 ),
[0058] Among them, features x1 to x 34 They correspond to detailed student health status indicators such as fever, no fever, sore throat, cough, chest symptoms, etc.
[0059] Furthermore, the feature vector is input into a preset number of machine learning models to obtain prediction results including:
[0060] The feature vectors were respectively input into the Bayesian model, decision tree model, support vector machine model, BP neural network model, logistic regression model and random forest model to obtain the prediction results of a single model.
[0061] Specifically, in the prior art, most disease prediction methods usually rely on a single machine learning model for prediction. However, the prediction ability of a single model is often limited by its own algorithm structure and is easily affected by specific data patterns, resulting in insufficient prediction accuracy and generalization ability. The present invention integrates a Bayesian model, a decision tree model, a support vector machine model, a BP neural network model, a logistic regression model, and a random forest model to form a multi-model integration method, which can give full play to the advantages of each model and make up for the shortcomings of a single model.
[0062] Beneficial effects: Through the weighted combination of multiple models, the present invention significantly improves the accuracy and robustness of the prediction results, and can better adapt to complex and heterogeneous student health data, thereby achieving efficient identification of multiple infectious diseases.
[0063] Furthermore, using the Bayesian model, the method for obtaining the prediction results is:
[0064]
[0065] Among them, C k is the kth disease category, X is the feature vector, P(X|C k ) is the likelihood of the feature under a specific disease, P(C k ) is the prior probability of the disease, and P(X) is the total probability distribution of the feature.
[0066] Specifically, based on Bayes' theorem, the model predicts diseases by maximizing the posterior probability.
[0067] Furthermore, using the decision tree model, the method for obtaining the prediction results is:
[0068]
[0069] Among them, Gain(X, A) is the information gain brought by feature X in data set A, H(X) represents the entropy of the data set, A is the partition feature, and X v is the subset after partitioning.
[0070] Specifically, a recursive partitioning method based on information gain or Gini coefficient is used to construct a decision tree.
[0071] Furthermore, the support vector machine model (SVM): Based on maximizing the classification boundary, the goal of SVM is to find the optimal hyperplane to classify the disease. Its optimization goal is:
[0072]
[0073] Among them, w is the normal vector of the hyperplane, b is the bias term, and y iis the label, x i are input features.
[0074] Furthermore, the BP neural network model: the neural network is trained by the back propagation algorithm, and the output of the neural network is the probability of the disease category. The loss function of the network is:
[0075]
[0076] Among them, y i is the actual disease label, is the predicted value of the neural network.
[0077] Furthermore, using the logistic regression model, the method for obtaining the prediction results is:
[0078]
[0079] Among them, w is the weight vector of the model, X is the input feature, and w T is the full-time weight vector, and y is the predicted target category.
[0080] Furthermore, the random forest model: By integrating multiple decision trees, the random forest classifies the disease. Its output is the average of the prediction results of all trees:
[0081]
[0082] Among them, f t Represents the prediction result of the tth decision tree.
[0083] Furthermore, weighted calculation is performed on the prediction results to obtain the optimal prediction result, including:
[0084] Perform weighted integration on the prediction results of individual models to obtain comprehensive prediction results;
[0085] The particle swarm optimization algorithm is used to optimize the weights of the comprehensive prediction results to obtain the best prediction results.
[0086] Specifically, model integration combines the prediction results of multiple models through the weighted average method and finally outputs a comprehensive prediction result. Assuming that the prediction results of each model are Y1, Y2, ..., Y6, and the weights of each model are w1, w2, ..., w6, the final integrated prediction result is:
[0087] Y=w1Y1+w2Y2+…+w6Y6
[0088] The weighted prediction results will be adjusted according to the weights optimized by the particle swarm algorithm to ensure the best prediction accuracy.
[0089] Traditional ensemble methods, such as simple voting or averaging strategies, fail to fully utilize the characteristics of each model. The present invention uses a weighted ensemble strategy to assign different weights according to the predictive ability of each model, which can more flexibly deal with the classification problems of different disease types.
[0090] Beneficial effects: The weighted strategy enables the method to dynamically adjust the contribution of each model according to actual needs when dealing with diseases with different characteristics, thereby improving the overall prediction effect, especially giving greater weight to some models that perform better in specific tasks, thereby enhancing the model's prediction ability for a variety of infectious diseases.
[0091] Furthermore, the method of optimizing the weight of the comprehensive prediction results using the particle swarm optimization algorithm is as follows:
[0092] v i+1 =wv i +c1r1(p i，best -w i )+c2r2(g best -w i )
[0093] w i+1 =w i +v i+1
[0094] Among them, w i is the weight vector of the ith particle, v i is the velocity of the ith particle, p i,best is the particle’s own historical optimal position, g best is the global optimal position, c1, c2 are learning factors, r1, r2 are random numbers, v i+1 is the velocity of the i+1th particle, w i+1 is the weight vector of the i+1th particle.
[0095] Specifically, in order to obtain the best model integration effect, the present invention uses the particle swarm optimization algorithm (PSO) to optimize the model weights. PSO searches for the optimal solution by simulating the movement of a particle swarm. Let the position of the i-th particle be w i =(w i1 , w i2 ..., w i6 ), with a velocity of v i .
[0096] In the prior art, weight allocation is often achieved through manual setting or simple optimization algorithms, and it is difficult to find the optimal model weight combination. The present invention optimizes the weights of each model by introducing a particle swarm optimization algorithm (PSO), which can effectively explore the global optimal solution and improve the rationality and accuracy of weight allocation.
[0097] Beneficial effects: The PSO algorithm has a strong global search capability and can quickly find the optimal weight combination, so that the prediction performance of the model integration reaches the best state. Compared with traditional optimization algorithms, particle swarm optimization converges faster and produces better results on complex problems, thereby improving the overall efficiency and accuracy of disease prediction.
[0098] The present invention also provides a campus potential infectious disease intelligent prediction machine learning integrated optimization system, including: a feature extraction module, a model prediction module, a model integration module and a weight optimization module;
[0099] A feature extraction module is used to obtain the health status data of students, extract several body features based on the health status data, and obtain a feature vector;
[0100] Model prediction module, used to input feature vectors into different types of machine learning models to obtain corresponding prediction results;
[0101] The model integration module is used to perform weighted calculation on the prediction results to obtain integrated prediction results;
[0102] The weight optimization module is used to optimize the weights of the model ensemble.
[0103] In summary, compared with the prior art, the present invention provides a more efficient, accurate and intelligent infectious disease prediction solution through multi-model integration, particle swarm optimization, comprehensive feature extraction and optimization design for campus scenarios. It has significant advantages in improving prediction accuracy, flexibility, automated processing and practical application effects, and provides strong technical support for the prevention and control of campus infectious diseases.
[0104] The above are only preferred specific implementations of the present application, but the protection scope of the present application 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 in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A machine learning integrated optimization method for intelligent prediction of potential infectious diseases on campus, characterized in that: include: Obtain student health status data; Extracting a number of body features based on the health status data to obtain a feature vector; Inputting the feature vector into different types of machine learning models to obtain corresponding prediction results; The corresponding prediction results are weighted to obtain the optimal prediction result.
2. According to claim 1, a method for intelligent prediction of potential infectious diseases on campus by machine learning integrated optimization is characterized in that: Inputting the feature vector into a preset number of machine learning models to obtain prediction results includes: The feature vectors are respectively input into the Bayesian model, decision tree model, support vector machine model, BP neural network model, logistic regression model and random forest model to obtain the prediction results of a single model.
3. According to claim 2, a method for intelligent prediction of potential infectious diseases on campus by machine learning integrated optimization is characterized in that: Using the Bayesian model, the method for obtaining the prediction result is: Among them, C k is the kth disease category, X is the feature vector, P(X|C k ) is the likelihood of the feature under a specific disease, P(C k ) is the prior probability of the disease, and P(X) is the total probability distribution of the feature.
4. According to claim 3, a method for intelligent prediction of potential infectious diseases on campus by machine learning integrated optimization is characterized in that: Using the decision tree model, the method for obtaining the prediction result is: Among them, Gain(X, A) is the information gain brought by feature X in data set A, H(X) represents the entropy of the data set, A is the partition feature, and X v is the subset after partitioning.
5. According to claim 3, a method for intelligent prediction of potential infectious diseases on campus by machine learning integrated optimization is characterized in that: Using the logistic regression model, the method for obtaining the prediction results is: Among them, w is the weight vector of the model, X is the input feature, and w T is the full-time weight vector, and y is the predicted target category.
6. The method for intelligent prediction of potential infectious diseases on campus by machine learning integrated optimization according to claim 2 is characterized in that: Performing weighted calculation on the prediction results to obtain the optimal prediction result includes: Performing weighted integration on the prediction results of the individual models to obtain a comprehensive prediction result; The particle swarm optimization algorithm is used to optimize the weight of the comprehensive prediction result to obtain the best prediction result.
7. The method for intelligent prediction of potential infectious diseases on campus by machine learning integrated optimization according to claim 6 is characterized in that: The method of optimizing the weight of the comprehensive prediction result by using the particle swarm optimization algorithm is as follows: u i+1 =wυ i +c1r1(p i,best -w i )+c2r2(g best —w i ) w i+1 =w i +υ i+1 Among them, w i is the weight vector of the ith particle, v i is the velocity of the ith particle, p i,best is the particle’s own historical optimal position, g best is the global optimal position, c1, c2 are learning factors, r1, r2 are random numbers, v i+1 is the velocity of the i+1th particle, w i+1 is the weight vector of the i+1th particle.
8. A campus potential infectious disease intelligent prediction machine learning integrated optimization system, characterized by: include: Feature extraction module, model prediction module, model integration module and weight optimization module; The feature extraction module is used to obtain the student's health status data, extract a number of body features based on the health status data, and obtain a feature vector; The model prediction module is used to input the feature vector into different types of machine learning models to obtain corresponding prediction results; The model integration module is used to perform weighted calculation on the prediction results to obtain integrated prediction results; The weight optimization module is used to optimize the weight of the model integration.