Intelligent infectious disease screening system for entry and exit personnel based on artificial intelligence

Abstract

The invention relates to the technical field of infectious disease screening, in particular to an entry and exit personnel infectious disease intelligent screening system based on artificial intelligence, which comprises a data acquisition subsystem, an artificial intelligence inspection subsystem, an intelligent screening subsystem and an efficient early warning subsystem, according to the scheme, dual detection is carried out through the artificial intelligence inspection subsystem and the intelligent screening subsystem, an infectious disease inspection model is constructed by using HGNN as a basic framework in the artificial intelligence inspection subsystem, two layers of multi-layer perceptron are added, and the feature extraction capability is enhanced; the prediction result of the LSTM and the pre-HGNN model is combined, high-risk individuals are further screened out, false alarm and missing alarm are reduced, a serow optimization algorithm and quantum calculation are introduced, a quantum superposition and interference principle is utilized to explore a plurality of parameter configurations, the LSTM, the quantum calculation and the serow optimization algorithm are combined, and the prediction result of the pre-HGNN model is obtained. And an efficient and intelligent infectious disease screening framework is formed.

Description

Technical Field

[0001] The present invention relates to the technical field of infectious disease screening, and specifically to an intelligent screening system for infectious diseases of entry-exit personnel based on artificial intelligence. Background Art

[0002] As infectious disease screening for people entering and leaving the country becomes increasingly stringent, general infectious disease screening systems only conduct one test based on the data and use general current models or shallow models. There are problems such as inaccurate test results and difficulty in dealing with complexity. There is no mechanism to screen out redundant features or noise, resulting in degraded model performance. For model optimization, when using traditional optimization methods, the efficiency is low in high-dimensional parameter space, resulting in a significant increase in optimization time, and there is a lack of dynamic adjustment capabilities, which makes it impossible to mine deep features in multimodal data, thereby reducing screening accuracy. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an intelligent screening system for infectious diseases of entry-exit personnel based on artificial intelligence. In view of the problems that general infectious disease screening systems have inaccurate detection results and are difficult to cope with complexity, the present invention performs dual detection through an artificial intelligence inspection subsystem and an intelligent screening subsystem, and uses HGNN as the infrastructure to construct an infectious disease inspection model in the artificial intelligence inspection subsystem, adds two layers of multi-layer perceptrons, enhances the feature extraction capability, and combines one-dimensional convolution and Dropout mechanisms to improve the generalization performance; the present invention uses LSTM combined with the prediction results of the front HGNN model, and further screens out individuals with high risks through dynamic analysis in the time dimension, reduces false positives and missed reports, introduces the spotted hyena optimization algorithm and quantum computing, and uses quantum superposition and interference principles to explore multiple parameter configurations, thereby improving the search efficiency of the global optimal solution, and combines LSTM, quantum computing and the spotted hyena optimization algorithm to form an efficient and intelligent infectious disease screening framework.

[0004] The technical solution adopted by the present invention is as follows: The intelligent screening system for infectious diseases of entry-exit personnel based on artificial intelligence provided by the present invention includes a data acquisition subsystem, an artificial intelligence inspection subsystem, an intelligent screening subsystem and an efficient early warning subsystem, specifically including the following contents:

[0005] The data collection subsystem collects physiological data, behavioral data and health record data of entry and exit personnel, physiological data includes body temperature, blood oxygen saturation and heart rate, behavioral data includes coughing, sneezing and breathing characteristics, and health records include infectious disease history, vaccination records and travel contact history;

[0006] The artificial intelligence inspection subsystem builds an infectious disease inspection model by using HGNN as the infrastructure, and conducts a comprehensive analysis of the physiological data, behavioral data, and health record data of incoming personnel to obtain prediction results;

[0007] The intelligent screening subsystem constructs an anomaly detection model based on LSTM, uses the anomaly detection model to screen the prediction results, and obtains the screening results;

[0008] The efficient early warning subsystem performs risk early warning and response according to the screening results of the intelligent screening subsystem.

[0009] Furthermore, in the artificial intelligence inspection subsystem, the infectious disease inspection model is used to conduct a comprehensive analysis to obtain an inspection report, which specifically includes the following steps:

[0010] Step S1: Data preprocessing: clustering the physiological data, behavioral data, and health record data of inbound and outbound personnel using the k-nearest neighbor method to obtain multimodal data;

[0011] Step S2: Model construction, adding two layers of multi-layer perceptron and attention mechanism on the basis of HGNN to build an infectious disease inspection model;

[0012] Step S3: Model training, collecting labeled historical data of infectious diseases, constructing a labeled training data set, and using the labeled training data set to train the infectious disease inspection model;

[0013] Step S4: Model optimization, optimizing the trained infectious disease inspection model, including hyperparameter optimization, regularization techniques, and cross-validation;

[0014] Step S5: Apply the model and use the optimized infectious disease inspection model to analyze the multimodal clustering of step S1 to obtain an inspection report.

[0015] Furthermore, in step S2, an infectious disease inspection model is constructed, which specifically includes the following contents:

[0016] Hypergraph construction module: constructs a multimodal hypergraph based on multimodal data;

[0017] Input layer: The input layer accepts the multimodal hypergraph generated by the hypergraph building module;

[0018] Feature extraction network: Extract key features from the multimodal hypergraph and add two layers of multilayer perceptrons, including the following:

[0019] The first layer of multi-layer perceptron: one-dimensional convolution, Sigmoid function and Dropout mechanism;

[0020] One-dimensional convolution of the first layer of multi-layer perceptron: input 1 channel, output 16 channels, stride 2;

[0021] The second layer of multi-layer perceptron: one-dimensional convolution, Sigmoid function and Dropout mechanism;

[0022] One-dimensional convolution of the second layer of multi-layer perceptron: input 16 channels, output 32 channels, stride 2;

[0023] Attention mechanism module: Generate channel weights through global average pooling and one-dimensional convolution, use Sigmoid activation function to adjust the weight of each channel, and output weighted key features;

[0024] Output layer: Use the Sigmoid function to map the weighted key features into probability distribution and output the prediction results.

[0025] Furthermore, in the intelligent screening subsystem, an anomaly detection model is constructed based on LSTM, and the prediction results are screened using the anomaly detection model, which specifically includes the following steps:

[0026] Step Q1: Data preparation, divide the prediction results into training set, validation set and test set in a ratio of 6:2:2;

[0027] Step Q2: Model construction, using the LSTM model as the basic architecture, introducing residual connections and attention mechanisms;

[0028] Step Q3: Model training, using the training set to train the screening model;

[0029] Step Q4: Model optimization, using the spotted hyena optimizer combined with quantum computing to optimize the parameters of the screening model;

[0030] Step Q5: Anomaly detection: Use the optimized screening model to detect the validation set, obtain the anomaly probability, set the probability threshold, divide the anomaly probability into normal and abnormal, and obtain the screening result.

[0031] Further, step Q4 specifically includes the following steps:

[0032] Step Q41: Qubit encoding and model parameter mapping, encoding the parameters of the screening model into quantum states, the value of each parameter is represented by the state of the qubit, mapped to the potential parameter space of the screening model, and a set of qubits is initialized to represent different parameter combinations;

[0033] Step Q42: Use a quantum rotating gate to update parameters. Use the quantum rotating gate operation to update the state of each quantum bit, use the quantum superposition principle to explore multiple parameter configurations, and approach a better parameter combination through interference effect;

[0034] Step Q43: Apply the spotted hyena optimization algorithm to perform fitness evaluation on the parameter combination represented by each group of qubits, obtain the evaluation results, use the spotted hyena optimization algorithm to update the evaluation results, and adjust the parameter combination corresponding to the qubits;

[0035] Step Q44: quantum bit update, adjusting the rotation angle and position of the quantum bit through a local optimization strategy according to the current fitness;

[0036] Step Q45: Optimize convergence, set the maximum number of iterations, iterate steps Q52 to Q54 until the maximum number of iterations is reached, and output the parameter combination of the final screening model.

[0037] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0038] (1) In view of the problems of inaccurate detection results and difficulty in coping with complexity in general infectious disease screening systems, the present invention performs dual detection through an artificial intelligence inspection subsystem and an intelligent screening subsystem, uses HGNN as the basic architecture to build an infectious disease inspection model in the artificial intelligence inspection subsystem, adds two layers of multi-layer perceptrons, enhances the feature extraction capability, and combines one-dimensional convolution and Dropout mechanisms to improve generalization performance;

[0039] (2) The present invention uses LSTM combined with the prediction results of the front-end HGNN model, and further screens out individuals with high risks through dynamic analysis in the time dimension, reducing false positives and negatives. It introduces the spotted hyena optimization algorithm and quantum computing, and uses the principles of quantum superposition and interference to explore multiple parameter configurations, thereby improving the search efficiency of the global optimal solution. The LSTM, quantum computing and spotted hyena optimization algorithm are combined to form an efficient and intelligent infectious disease screening framework. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of the artificial intelligence-based intelligent screening system for infectious diseases for entry-exit personnel proposed by the present invention;

[0041] Figure 2 A schematic diagram of the process of the method for comprehensive analysis using the infectious disease inspection model proposed in the present invention;

[0042] Figure 3 A flow chart of the method proposed in the present invention using the spotted hyena optimizer combined with quantum computing.

[0043] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] Example 1, see Figure 1 The artificial intelligence-based intelligent screening system for infectious diseases of entry-exit personnel provided by the present invention includes a data acquisition subsystem, an artificial intelligence inspection subsystem, an intelligent screening subsystem and an efficient early warning subsystem, and specifically includes the following contents:

[0046] The data collection subsystem collects physiological data, behavioral data and health record data of entry and exit personnel, physiological data includes body temperature, blood oxygen saturation and heart rate, behavioral data includes coughing, sneezing and breathing characteristics, and health records include infectious disease history, vaccination records and travel contact history;

[0047] The artificial intelligence inspection subsystem builds an infectious disease inspection model by using HGNN as the infrastructure, and conducts a comprehensive analysis of the physiological data, behavioral data, and health record data of incoming personnel to obtain prediction results;

[0048] The intelligent screening subsystem constructs an anomaly detection model based on LSTM, uses the anomaly detection model to screen the prediction results, and obtains the screening results;

[0049] The efficient early warning subsystem performs risk early warning and response according to the screening results of the intelligent screening subsystem.

[0050] Example 2, see Figure 2 This embodiment is based on the above embodiment. In the artificial intelligence inspection subsystem, an infectious disease inspection model is used to perform comprehensive analysis to obtain an inspection report. Specifically, the following steps are included:

[0051] Step S1: Data preprocessing: clustering the physiological data, behavioral data, and health record data of inbound and outbound personnel using the k-nearest neighbor method to obtain multimodal data;

[0052] Step S2: Model construction, adding two layers of multi-layer perceptron and attention mechanism on the basis of HGNN to build an infectious disease inspection model;

[0053] Step S3: Model training, collecting labeled historical data of infectious diseases, constructing a labeled training data set, and using the labeled training data set to train the infectious disease inspection model;

[0054] Step S4: Model optimization, optimizing the trained infectious disease inspection model, including hyperparameter optimization, regularization techniques, and cross-validation;

[0055] Step S5: Apply the model and use the optimized infectious disease inspection model to analyze the multimodal clustering of step S1 to obtain an inspection report.

[0056] Embodiment 3: This embodiment is based on the above embodiment. Step S2 is to construct an infectious disease inspection model, which specifically includes the following contents:

[0057] Hypergraph construction module: constructs a multimodal hypergraph based on multimodal data;

[0058] Input layer: The input layer accepts the multimodal hypergraph generated by the hypergraph building module;

[0059] Feature extraction network: Extract key features from the multimodal hypergraph and add two layers of multilayer perceptrons, including the following:

[0060] The first layer of multi-layer perceptron: one-dimensional convolution, Sigmoid function and Dropout mechanism;

[0061] One-dimensional convolution of the first layer of multi-layer perceptron: input 1 channel, output 16 channels, stride 2;

[0062] The second layer of multi-layer perceptron: one-dimensional convolution, Sigmoid function and Dropout mechanism;

[0063] One-dimensional convolution of the second layer of multi-layer perceptron: input 16 channels, output 32 channels, stride 2;

[0064] Attention mechanism module: Generate channel weights through global average pooling and one-dimensional convolution, use Sigmoid activation function to adjust the weight of each channel, and output weighted key features;

[0065] Output layer: Use the Sigmoid function to map the weighted key features into probability distribution and output the prediction results.

[0066] Embodiment 4: This embodiment is based on the above embodiment. In the intelligent screening subsystem, an anomaly detection model is constructed based on LSTM, and the prediction results are screened using the anomaly detection model. Specifically, the following steps are included:

[0067] Step Q1: Data preparation, divide the prediction results into training set, validation set and test set in a ratio of 6:2:2;

[0068] Step Q2: Model construction, using the LSTM model as the basic architecture, introducing residual connections and attention mechanisms;

[0069] Step Q3: Model training, using the training set to train the screening model;

[0070] Step Q4: Model optimization, using the spotted hyena optimizer combined with quantum computing to optimize the parameters of the screening model;

[0071] Step Q5: Anomaly detection: Use the optimized screening model to detect the validation set, obtain the anomaly probability, set the probability threshold, divide the anomaly probability into normal and abnormal, and obtain the screening result.

[0072] Example 5, see Figure 3 This embodiment is based on the above embodiment, step Q4, specifically includes the following steps:

[0073] Step Q41: Qubit encoding and model parameter mapping, encoding the parameters of the screening model into quantum states, the value of each parameter is represented by the state of the qubit, mapped to the potential parameter space of the screening model, and a set of qubits is initialized to represent different parameter combinations;

[0074] Step Q42: Use a quantum rotating gate to update parameters. Use the quantum rotating gate operation to update the state of each quantum bit, use the quantum superposition principle to explore multiple parameter configurations, and approach a better parameter combination through interference effect;

[0075] Step Q43: Apply the spotted hyena optimization algorithm to perform fitness evaluation on the parameter combination represented by each group of qubits, obtain the evaluation results, use the spotted hyena optimization algorithm to update the evaluation results, and adjust the parameter combination corresponding to the qubits;

[0076] Step Q44: quantum bit update, adjusting the rotation angle and position of the quantum bit through a local optimization strategy according to the current fitness;

[0077] Step Q45: Optimize convergence, set the maximum number of iterations, iterate steps Q52 to Q54 until the maximum number of iterations is reached, and output the parameter combination of the final screening model.

[0078] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0079] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0080] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. An intelligent screening system for infectious diseases of entry-exit personnel based on artificial intelligence, characterized by: It includes data collection subsystem, artificial intelligence inspection subsystem, intelligent screening subsystem and efficient early warning subsystem, specifically including the following contents: The data collection subsystem collects physiological data, behavioral data and health record data of entry and exit personnel, physiological data includes body temperature, blood oxygen saturation and heart rate, behavioral data includes coughing, sneezing and breathing characteristics, and health records include infectious disease history, vaccination records and travel contact history; The artificial intelligence inspection subsystem builds an infectious disease inspection model by using HGNN as the infrastructure, and conducts a comprehensive analysis of the physiological data, behavioral data, and health record data of incoming personnel to obtain prediction results; The intelligent screening subsystem constructs an anomaly detection model based on LSTM, uses the anomaly detection model to screen the prediction results, and obtains the screening results; The efficient early warning subsystem performs risk early warning and response according to the screening results of the intelligent screening subsystem.

2. The artificial intelligence-based intelligent screening system for infectious diseases of entry-exit personnel according to claim 1 is characterized by: In the artificial intelligence inspection subsystem, the infectious disease inspection model is used to conduct a comprehensive analysis to obtain an inspection report, which specifically includes the following steps: Step S1: Data preprocessing: clustering the physiological data, behavioral data, and health record data of inbound and outbound personnel using the k-nearest neighbor method to obtain multimodal data; Step S2: Model construction, adding two layers of multi-layer perceptron and attention mechanism on the basis of HGNN to build an infectious disease inspection model; Step S3: Model training, collecting labeled historical data of infectious diseases, constructing a labeled training data set, and using the labeled training data set to train the infectious disease inspection model; Step S4: Model optimization, optimizing the trained infectious disease inspection model, including hyperparameter optimization, regularization techniques, and cross-validation; Step S5: Apply the model and use the optimized infectious disease inspection model to analyze the multimodal clustering of step S1 to obtain an inspection report.

3. The artificial intelligence-based intelligent screening system for infectious diseases of entry-exit personnel according to claim 2 is characterized by: Construct an infectious disease inspection model, which includes the following: Hypergraph construction module: constructs a multimodal hypergraph based on multimodal data; Input layer: The input layer accepts the multimodal hypergraph generated by the hypergraph building module; Feature extraction network: Extract key features from the multimodal hypergraph and add two layers of multi-layer perceptrons, including the following: The first layer of multi-layer perceptron: one-dimensional convolution, Sigmoid function and Dropout mechanism; One-dimensional convolution of the first layer of multi-layer perceptron: input 1 channel, output 16 channels, stride 2; The second layer of multi-layer perceptron: one-dimensional convolution, Sigmoid function and Dropout mechanism; One-dimensional convolution of the second layer of multi-layer perceptron: input 16 channels, output 32 channels, stride 2; Attention mechanism module: Generate channel weights through global average pooling and one-dimensional convolution, use Sigmoid activation function to adjust the weight of each channel, and output weighted key features; Output layer: Use the Sigmoid function to map the weighted key features into probability distribution and output the prediction results.

4. The artificial intelligence-based intelligent screening system for infectious diseases of entry-exit personnel according to claim 1 is characterized by: In the intelligent screening subsystem, an anomaly detection model is constructed based on LSTM, and the prediction results are screened using the anomaly detection model, which specifically includes the following steps: Step Q1: Data preparation, divide the prediction results into training set, validation set and test set in a ratio of 6:2:2; Step Q2: Model construction, using the LSTM model as the basic architecture, introducing residual connections and attention mechanisms; Step Q3: Model training, using the training set to train the screening model; Step Q4: Model optimization, using the spotted hyena optimizer combined with quantum computing to optimize the parameters of the screening model; Step Q5: Anomaly detection: Use the optimized screening model to detect the validation set, obtain the anomaly probability, set the probability threshold, divide the anomaly probability into normal and abnormal, and obtain the screening result.

Images