Infectious Disease Data Processing Method, Device, Computer, and Storage Medium

Abstract

This application relates to the fields of artificial intelligence and digital medicine, and discloses an infectious disease data processing method, device, computer device, and storage medium, including: obtaining infectious disease data to be processed; preprocessing the infectious disease data according to a preset data dimension to generate a feature matrix; inputting the feature matrix into a preset infectious disease prediction model, and the infectious disease prediction model performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix; wherein, the infectious disease prediction model is pre-trained to a convergence state according to an unlabeled sample model for extracting hidden layer information of input data; generating a disease index of the infectious disease data according to the hidden layer features. Writing the disease index into a preset Cartesian coordinate system to generate a schematic diagram of the representation result of the infectious disease data, and inputting the schematic diagram of the representation result into a preset infectiousness recognition model to generate an infectiousness classification result of the infectious disease data.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and particularly to an infectious disease data processing method, a system, and a computer storage medium. Background Art

[0002] The infectious disease index is used for infectious disease early warning, and its function is similar to the "weather forecast" in the field of infectious diseases, indicating the risk of recent infectious diseases, which can help relevant agencies better prevent and manage infectious disease outbreaks and reduce the harm caused by them. Taking the infectious disease index as an example, the infectious disease index is based on data collected by an epidemic monitoring network, such as multi-dimensional influenza monitoring data such as the percentage of influenza-like cases, the number of influenza outbreaks, and the number of influenza cases, and integrates them into an infectious disease index.

[0003] In the current method for integrating the infectious disease index, first, weights are determined for the data of each dimension through expert experience, and then the data of each dimension are linearly summed according to the weights. However, there are certain problems with the existing method:

[0004] 1. The monitoring quality of the data of each dimension in different regions is different, and there are also spatial differences in the scale and occurrence characteristics of infectious diseases. In addition, the experience of experts is too subjective and cannot be transferred, so the accuracy and adaptability of the infectious disease index are poor.

[0005] 2. The current method is to simply linearly weight and average the data of each dimension, resulting in the information covered by the infectious disease index in different periods being fixed and unable to adapt to different influenza outbreak situations (such as different dimensions of monitoring data should be concerned in the influenza epidemic season and the non-epidemic season), so the accuracy of the infectious disease index is poor. Summary of the Invention

[0006] The purpose of the present invention is to provide an infectious disease data processing method, a system, and a computer storage medium to at least solve the problem of poor accuracy of the infectious disease index obtained by the existing infectious disease index integration method.

[0007] To solve the above technical problems, the present invention provides an infectious disease data processing method, including:

[0008] Obtaining the infectious disease data to be processed;

[0009] Preprocessing the infectious disease data according to a preset data dimension to generate a feature matrix;

[0010] Inputting the feature matrix into a preset infectious disease prediction model, and the infectious disease prediction model performs downsampling processing on the feature matrix to generate a hidden layer feature representing the feature matrix; wherein, the infectious disease prediction model is pre-trained to a convergence state according to an unlabeled sample model and is used to extract the hidden layer information of the input data;

[0011] Generate the disease index of the infectious disease data based on the hidden layer features;

[0012] Write the disease index into a preset Cartesian coordinate system to generate a disease index graph of the infectious disease data, and input the disease index graph into a preset infectious classification model to generate an infectious classification result of the infectious disease data.

[0013] Optionally, in the infectious disease data processing method, the method for obtaining the infectious disease data to be processed includes:

[0014] Obtain a preset acquisition time sequence;

[0015] Collect multiple groups of original monitoring data in sequence according to the acquisition time sequence within a preset detection duration;

[0016] Arrange the multiple groups of original monitoring data in ascending order according to the acquisition order to generate infectious disease data.

[0017] Optionally, in the infectious disease data processing method, the method for preprocessing the infectious disease data according to a preset data dimension to generate a feature matrix includes:

[0018] Read a preset data dimension;

[0019] Extract data from the infectious disease data in sequence according to the data dimension to generate multiple groups of time series vectors corresponding to the infectious disease data;

[0020] Arrange the multiple groups of time series vectors in sequence according to the corresponding extraction time sequence to generate a feature matrix.

[0021] Optionally, in the infectious disease data processing method, the method for inputting the feature matrix into a preset infectious disease prediction model, where the infectious disease prediction model performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix includes:

[0022] Call the weight matrix of the infectious disease prediction model, where the dimension of the weight matrix is lower than the dimension of the feature matrix;

[0023] Input the feature matrix into the infectious disease prediction model, and perform an inner product operation on the weight matrix and the feature matrix to generate hidden layer features.

[0024] Optionally, in the infectious disease data processing method, the method for generating the disease index of the infectious disease data based on the hidden layer features includes:

[0025] Statistically analyze the data represented by the hidden layer features according to the dimension of the weight matrix to generate a multi-dimensional array;

[0026] Display the multi-dimensional array in a preset infectious disease index coordinate system to generate a disease index graph, where arrays of different dimensions are distinguished by display colors and / or shapes.

[0027] Optionally, in the infectious disease data processing method, after generating the disease index of the infectious disease data according to the hidden layer features, the infectious disease data processing method further includes:

[0028] Input the disease index graph into a preset infectious classification model, where the infectious classification model is a neural network model trained to a convergent state by a supervised training method for risk classification of infectious diseases;

[0029] Read the risk classification result output by the infectious classification model according to the disease index graph;

[0030] Based on the classification result, match a response strategy corresponding to the classification result in a preset response database.

[0031] Optionally, the infectious disease prediction model includes an encoder and a decoder, and the training method of the infectious disease prediction model includes:

[0032] Collect a training sample set, where the training sample set includes multiple training samples, and each training sample includes a set of historical infectious disease data;

[0033] Input each training sample into the encoder in sequence, so that the encoder performs downsampling processing on the training sample to generate sample hidden layer features;

[0034] Input the sample hidden layer features into the decoder, so that the decoder performs upsampling processing on the sample hidden layer features to generate sample restored data;

[0035] Calculate the feature distance between the sample restored data and the historical infectious disease data corresponding to the training sample according to a preset loss function;

[0036] Compare the feature distance with a preset distance threshold. When the feature distance is greater than the distance threshold, repeatedly iterate to adjust the weight parameters of the encoder according to the feature distance until the feature distance is less than or equal to the feature threshold.

[0037] To solve the above technical problems, an embodiment of the present invention further provides an infectious disease data processing device, including:

[0038] An acquisition module, configured to acquire infectious disease data to be processed;

[0039] A processing module for preprocessing the infectious disease data according to a preset data dimension to generate a feature matrix;

[0040] A classification module for inputting the feature matrix into a preset infectious disease prediction model, which performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix; wherein, the infectious disease prediction model is pre-trained to a convergence state according to an unlabeled sample model for extracting hidden layer information of input data;

[0041] An execution module for generating a disease index of the infectious disease data according to the hidden layer features;

[0042] An identification module for writing the disease index into a preset Cartesian coordinate system to generate a disease index map of the infectious disease data, and inputting the disease index map into a preset infectious classification model to generate an infectious classification result of the infectious disease data.

[0043] Optionally, the infectious disease data processing device further includes:

[0044] A first acquisition sub-module for acquiring a preset acquisition time sequence;

[0045] A first acquisition sub-module for sequentially acquiring multiple groups of original monitoring data according to the acquisition time sequence within a preset detection duration;

[0046] A first sorting sub-module for sorting multiple groups of the original monitoring data in ascending order according to the acquisition order to generate infectious disease data.

[0047] Optionally, the infectious disease data processing device further includes:

[0048] A first reading sub-module for reading a preset data dimension;

[0049] A first extraction sub-module for sequentially extracting data from the infectious disease data according to the data dimension to generate multiple groups of time series vectors corresponding to the infectious disease data;

[0050] A second sorting sub-module for sorting multiple groups of the time series vectors in sequence according to the corresponding extraction time sequence to generate a feature matrix.

[0051] Optionally, the infectious disease data processing device further includes:

[0052] A first calling sub-module for calling a weight matrix of an infectious disease prediction model, wherein the dimension of the weight matrix is lower than the dimension of the feature matrix;

[0053] The first operator module is used to input the feature matrix into the infectious disease prediction model, so that the weight matrix and the feature matrix perform an inner product operation to generate hidden layer features.

[0054] Optionally, the infectious disease data processing device further includes:

[0055] The first statistical sub-module is used to statistically process the data represented by the hidden layer features according to the dimensions of the weight matrix to generate a multi-dimensional array;

[0056] The first generation sub-module is used to display the multi-dimensional array in a preset infectious disease index coordinate system to generate a disease index map, where arrays of different dimensions are distinguished by display colors and / or shapes.

[0057] Optionally, the infectious disease data processing device further includes:

[0058] The first input sub-module is used to input the disease index map into a preset infectious classification model, where the infectious classification model is a neural network model trained to a convergent state by a supervised training method for risk classification of infectious diseases;

[0059] The second reading sub-module is used to read the risk classification result output by the infectious classification model according to the disease index map;

[0060] The first matching sub-module is used to match a response strategy corresponding to the classification result in a preset response database based on the classification result.

[0061] Optionally, the infectious disease data processing device further includes:

[0062] The second acquisition sub-module is used to acquire a training sample set, where the training sample set includes multiple training samples, and each training sample includes a set of historical infectious disease data;

[0063] The second input sub-module is used to sequentially input each training sample into the encoder, so that the encoder performs downsampling processing on the training sample to generate sample hidden layer features;

[0064] The third input sub-module is used to input the sample hidden layer features into the decoder, so that the decoder performs upsampling processing on the sample hidden layer features to generate sample restored data;

[0065] The second operator module is used to calculate the feature distance between the sample restored data and the historical infectious disease data corresponding to the training sample according to a preset loss function;

[0066] The first iterative sub-module is used to compare the feature distance with a preset distance threshold. When the feature distance is greater than the distance threshold, the weight parameters of the encoder are iteratively adjusted according to the feature distance until the feature distance is less than or equal to the feature threshold.

[0067] To solve the above technical problems, an embodiment of the present invention further provides a computer device, including a memory and a processor. A computer-readable instruction is stored in the memory. When the computer-readable instruction is executed by the processor, the processor executes the steps of the above-mentioned infectious disease data processing method.

[0068] To solve the above technical problems, an embodiment of the present invention further provides a storage medium storing a computer-readable instruction. When the computer-readable instruction is executed by one or more processors, the one or more processors execute the steps of the above-mentioned infectious disease data processing method.

[0069] The infectious disease data processing method, system, and computer storage medium provided by the present invention include: obtaining infectious disease data to be processed; preprocessing the infectious disease data according to a preset data dimension to generate a feature matrix; inputting the feature matrix into a preset infectious disease prediction model, and the infectious disease prediction model performs downsampling processing on the feature matrix to generate a hidden layer feature representing the feature matrix; wherein, the infectious disease prediction model is pre-trained based on an unlabeled sample model to a convergence state and is used to extract the hidden layer information of the input data; generating a disease index of the infectious disease data according to the hidden layer feature. By using the trained infectious disease prediction model for processing and analysis, the infectious disease index can be automatically calculated based on the regional real data, so it can be applied to the infectious disease index with spatial differences; secondly, the downsampling processing process of the infectious disease prediction model is a complex non-linear process, which can more accurately fit the relationship between different-dimensional monitoring matrices. Therefore, the infectious disease index obtained by the model can cover the monitoring data of each dimension as much as possible and adapt to different infectious disease outbreak scenarios. It solves the problem that the accuracy of the infectious disease index obtained by the existing infectious disease index integration method is poor. By visualizing the obtained disease index and performing image recognition on the visualized graph, the infectious disease classification result representing the infectiousness of the infectious disease data is obtained, which can enable the infectious disease classification model to focus on the core indicators of the infectious disease data and make the infectious disease classification result more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The above and / or additional aspects and advantages of the present application will become apparent and easy to understand from the following description of the embodiments in conjunction with the drawings, where:

[0071] Figure 1Schematic diagram of the basic process of the infectious disease data processing method for a specific embodiment of the present application;

[0072] Figure 2 Schematic diagram of the structure of the infectious disease prediction model for a specific embodiment of the present application;

[0073] Figure 3 Disease index graph of the infectious disease prediction model for a specific embodiment of the present application;

[0074] Figure 4 Schematic diagram of the acquisition process of the infectious disease data for a specific embodiment of the present application;

[0075] Figure 5 Schematic diagram of the process of generating the feature matrix for a specific embodiment of the present application;

[0076] Figure 6 Schematic diagram of the process of calculating the hidden layer features for a specific embodiment of the present application;

[0077] Figure 7 Schematic diagram of the process of generating the disease index graph for a specific embodiment of the present application;

[0078] Figure 8 Schematic diagram of the process of risk classification for the disease index graph for a specific embodiment of the present application;

[0079] Figure 9 Schematic diagram of the training process of the infectious disease prediction model for a specific embodiment of the present application;

[0080] Figure 10 Schematic diagram of the basic structure of the infectious disease data processing device for an embodiment of the present application;

[0081] Figure 11 Basic structural block diagram of a computer device for an embodiment of the present application. Detailed Description of the Embodiment

[0082] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be construed as a limitation of the present application.

[0083] Those skilled in the art can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of this application means the presence of the stated features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups.

[0084] Those skilled in the art can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0085] Those skilled in the art can understand that the "terminal" used herein includes both a device with a wireless signal receiver, which only has a wireless signal receiver without transmission ability, and a device with receiving and transmitting hardware, which has receiving and transmitting hardware capable of performing two-way communication on a two-way communication link. Such devices may include: cellular or other communication devices, which have a single-line display or a multi-line display or a cellular or other communication device without a multi-line display; PCS (Personal Communications Service), which can combine voice, data processing, fax and / or data communication capabilities; PDA (Personal Digital Assistant), which may include a radio frequency receiver, a pager, Internet / intranet access, a web browser, a notepad, a calendar and / or a GPS (Global Positioning System) receiver; conventional laptop and / or palm-top computers or other devices, which are conventional laptop and / or palm-top computers or other devices with and / or including a radio frequency receiver. The "terminal" used herein can be portable, transportable, installed in a vehicle (air, sea and / or land), or suitable for and / or configured to operate locally, and / or operate in a distributed form at any other location on the earth and / or in space. The "terminal" used herein can also be a communication terminal, an Internet access terminal, a music / video playback terminal, such as a PDA, a MID (Mobile Internet Device) and / or a mobile phone with music / video playback function, or can also be devices such as a smart TV, a set-top box, etc.

[0086] Please refer toFigure 1 , Figure 1 is a schematic diagram of the basic process of the infectious disease data processing method in this embodiment.

[0087] As Figure 1 shown, an infectious disease data processing method includes:

[0088] S1100. Obtain the infectious disease data to be processed;

[0089] In this embodiment, the infectious disease data within the jurisdiction is collected at equal intervals by means of regular collection.

[0090] The jurisdiction refers to the implementation scope of the infectious disease data processing method in this embodiment. For example, the infectious disease data of medical institutions within a hospital, county, city, province or country. In this embodiment, the infectious disease data has M dimensions. The M-dimensional infectious disease data includes, but is not limited to, the percentage of infectious disease samples, the number of infectious disease outbreaks, the number of infectious disease cases, and the positive rate of infectious disease viruses. For example, when obtaining the infectious disease index of influenza, the percentage of influenza-like cases, the number of influenza outbreaks, the number of influenza cases, and the positive rate of influenza viruses can be obtained. That is, the influenza surveillance data is 4-dimensional, and at this time M = 4 in the matrix.

[0091] The infectious disease data is composed of data at n moments. After each group of data is collected, each group of data is sorted according to the collection time. After n groups of data are collected, the infectious disease data is generated.

[0092] S1200. Preprocess the infectious disease data according to the preset data dimensions to generate a feature matrix;

[0093] After the infectious disease data is collected, it is classified and statistically analyzed according to the preset data dimensions. For example, data extraction is performed on the infectious disease data in data dimensions such as the percentage of infectious disease samples, the number of infectious disease outbreaks, the number of infectious disease cases, and the positive rate of infectious disease viruses. It should be noted that the dimensions of classification and statistics are not limited to this. According to the different specific application scenarios, in some embodiments, the statistical dimensions of the data can be any dimension data set by the user. The process of data extraction is the process of preprocessing the infectious disease data.

[0094] In the process of data statistical extraction, m-dimensional data is extracted from the infectious disease data, where M is greater than or equal to m. Therefore, a feature matrix of n * m dimensions can be obtained through data statistics.

[0095] S1300. Input the feature matrix into a preset infectious disease prediction model. The infectious disease prediction model performs downsampling on the feature matrix to generate hidden layer features representing the feature matrix. Among them, the infectious disease prediction model is pre-trained to a convergent state according to an unlabeled sample model and is used to extract hidden layer information of input data.

[0096] After the feature matrix is generated, input the feature matrix into a preset infectious disease prediction model. As Figure 2 shown, in this embodiment, the infectious disease prediction model is constructed based on an autoencoder neural network. The construction of the infectious disease prediction model includes: an input layer Xm, an encoder g, a hidden layer, a decoder f, and an output layer that are sequentially structured. Input the m*n feature matrix into the input layer Xm of the infectious disease prediction model. The encoder g downsamples the feature matrix and maps the feature matrix within a preset low-dimensional space, which is equivalent to compressing the data of the feature matrix. After being encoded by the encoder g and mapped to the hidden layer, a representation feature Z is obtained. The representation feature Z is the hidden layer feature of the feature matrix.

[0097] The training process of the infectious disease prediction model is an unsupervised training process. The sample data used for training does not need to be labeled. After being encoded by the encoder g and mapped to the hidden layer, a representation feature Z is obtained, where the mapping process is denoted as g(Xm); the decoder f is used to map the representation feature Z back to the original feature space to obtain a reconstructed sample, and the reconstructed sample is output through the output layer, where the decoding process is denoted as f(Z). Then, calculate the feature distance between the reconstructed sample and the training sample, and correct the weights of the encoder g based on the feature distance until the infectious disease prediction model is trained to convergence.

[0098] S1400. Generate a disease index for the infectious disease data according to the hidden layer features.

[0099] Calculate the disease index for the infectious disease data based on the hidden layer features. For example, perform weighted summation on each of the m-dimensional data in the hidden layer features to generate m types of disease indices.

[0100] As Figure 3 shown, it shows the output results corresponding to the input of the infectious disease prediction model with a time series length n of 150. It can be seen that the output results of the infectious disease prediction model are relatively evenly distributed in the hidden layer representation space, reflecting the differential representation results.

[0101] S1500. Write the disease index into a preset Cartesian coordinate system to generate a disease index graph of the infectious disease data, and input the disease index graph into a preset infectiousness classification model to generate an infectiousness classification result of the infectious disease data.

[0102] Write the disease index into a preset Cartesian coordinate system to generate a disease index graph of infectious disease data. As Figure 3 shown, arrange the generated disease indices in chronological order and write them into the Cartesian coordinate system in sequence to generate coordinate points, and a disease index graph of infectious disease data is composed of multiple coordinate points.

[0103] Input the disease index graph into a preset infectiousness classification model. The infectiousness classification model is a neural network model that has been pre-trained to a convergent state for classifying the infectiousness of infectious diseases according to image data. The infectiousness classification model can be constructed by (but not limited to) one of a convolutional neural network model, a deep convolutional neural network model, or a recurrent neural network model, or a variant model of the above models, and is trained to a convergent state through supervised training, and can generate an infectiousness classification result of infectious disease data according to the disease index graph.

[0104] The infectiousness classification result has three classification results: high, medium, and low. When the infectiousness classification result is high, the system sends an alarm message in the form of text, music, or image / video.

[0105] Obtain the infectious disease data to be processed; preprocess the infectious disease data according to preset data dimensions to generate a feature matrix; input the feature matrix into a preset infectious disease prediction model. The infectious disease prediction model performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix. Among them, the infectious disease prediction model is pre-trained to a convergent state according to an unlabeled sample model and is used to extract hidden layer information of input data; generate the disease index of the infectious disease data according to the hidden layer features. By using the trained infectious disease prediction model for processing and analysis, the calculation of the infectious disease index can be automatically based on regional real data, so it can be applied to infectious disease indices with spatial differences; secondly, the downsampling process of the infectious disease prediction model is a complex non-linear process, which can more accurately fit the relationship between monitoring matrices of different dimensions, so that the infectious disease index obtained by the model can cover as much monitoring data of each dimension as possible and adapt to different infectious disease outbreak scenarios. It solves the problem that the accuracy of the infectious disease index obtained by the existing infectious disease index integration method is poor. By visualizing the obtained disease index and performing image recognition on the visualized graph, an infectiousness classification result representing the infectious disease of the infectious disease data is obtained, which can enable the infectiousness classification model to focus on the core indicators of the infectious disease data and make the infectiousness classification result more accurate.

[0106] In some embodiments, the infectious disease data is composed of multiple groups of detection data arranged in chronological order. Please refer to Figure 4 , Figure 4 which is a schematic diagram of the acquisition process of the infectious disease data in this embodiment.

[0107] AsFigure 4 As shown in Figure 4 , S1100 includes:

[0108] S1111. Obtain a preset acquisition time sequence;

[0109] In this embodiment, the infectious disease data is aggregated from the detection data within the affiliated region. Each group of detection data has a fixed acquisition time sequence. For example, a group of detection data is acquired every 1 hour. However, the acquisition time sequence of the detection data is not limited to this. According to different specific application scenarios, in some embodiments, the duration of the acquisition time sequence is 2 hours, 5 hours, 18 hours, 24 hours, one week, or one month, etc., which is a custom duration suitable for the scenario requirements.

[0110] S1112. Acquire multiple groups of original monitoring data in sequence according to the acquisition time sequence within a preset detection duration;

[0111] According to the determined acquisition time sequence, obtain n groups of original monitoring data within the set detection time. The preset detection duration is a measurement period, and the setting of the detection duration is related to the required original monitoring data. For example, if it is set that the infectious disease data includes 10 groups of original monitoring data, then the detection duration is 9 * the acquisition time sequence. When the set infectious disease data includes C groups of original monitoring data, the detection duration is: (C - 1) * the acquisition time sequence.

[0112] S1113. Arrange the multiple groups of the original monitoring data in ascending order according to the acquisition order to generate infectious disease data.

[0113] According to the acquisition time of the multiple groups of original monitoring data, arrange the multiple groups of the original monitoring data in ascending order according to the sequence of time, and arrange the original monitoring data with an earlier acquisition time before other original monitoring data to generate n groups of original monitoring data sorted by acquisition time. This original monitoring data sorted by acquisition time is the infectious disease data.

[0114] In some embodiments, after the infectious disease data is acquired, it is necessary to perform data statistics on the infectious disease data to generate a feature matrix. Please refer to Figure 5 , Figure 5 which is the schematic flowchart of generating the feature matrix for this embodiment.

[0115] As Figure 5 shown in Figure 5 , S1200 includes:

[0116] S1211. Read a preset data dimension;

[0117] In this embodiment, the infectious disease data has M dimensions. The M-dimensional infectious disease data includes, but is not limited to, the percentage of infectious disease samples, the number of infectious disease outbreaks, the number of infectious disease cases, and the positive rate of infectious disease viruses. For example, when obtaining the infectious disease index of influenza, the percentage of influenza-like cases, the number of influenza outbreaks, the number of influenza cases, and the positive rate of influenza viruses can be obtained.

[0118] Classify and count the infectious disease data according to the preset data dimensions. For example, extract the infectious disease data using data dimensions such as the percentage of infectious disease samples, the number of infectious disease outbreaks, the number of infectious disease cases, and the positive rate of infectious disease viruses. It should be noted that the dimensions of the classification and statistics are not limited to this. According to different specific application scenarios, in some embodiments, the statistical dimensions of the data can be any dimension data set by the user.

[0119] S1212. Extract the data from the infectious disease data in sequence according to the data dimensions to generate multiple groups of time series vectors corresponding to the infectious disease data.

[0120] After obtaining the data dimensions, extract the data in the infectious disease data according to each dimension in the data dimensions. The extraction method is as follows: Extract the data with the same dimension in the infectious disease data to generate a group of time series vectors for this dimension. The result of the data extraction is that each data dimension corresponds to a generated group of time series vectors, so multiple data dimensions correspond to multiple groups of time series vectors.

[0121] S1213. Sort the multiple groups of time series vectors in sequence according to the corresponding extraction time sequence to generate a feature matrix.

[0122] The extraction times of the data have a sequence relationship. Sort the time series vectors in sequence according to the extraction time sequence corresponding to each group to generate a feature matrix. For example, in the process of data statistics and extraction, an m-dimensional time series vector is extracted from the infectious disease data. Since the infectious disease data is composed of n groups of data, a feature matrix with an n*m dimension can be obtained through data statistics.

[0123] In some embodiments, the infectious disease prediction model includes an encoder, and the encoder includes a weight matrix. Calculate the hidden layer features based on this weight matrix for the feature matrix. Please refer to Figure 6 , Figure 6 This is the flow diagram for calculating the hidden layer features in this embodiment.

[0124] As Figure 6 shown, S1300 includes:

[0125] S1311. Invoke the weight matrix of the infectious disease prediction model, where the dimension of the weight matrix is lower than the dimension of the feature matrix.

[0126] In this embodiment, the infectious disease prediction model includes an encoder, and the encoder includes a set of weight matrices. This set of weight matrices is equivalent to a set of data filters, which can amplify the useful data in the feature data, making the key data prominent, while shrinking or zeroing the unimportant data and weakening the unimportant data. The processes of strengthening and weakening are carried out simultaneously, which can make the key data in the feature data prominent and is conducive to improving the accuracy of index calculation.

[0127] The dimension of the weight matrix is lower than that of the feature matrix. Therefore, through the weight matrix, the feature matrix can be mapped into a space with the same dimension as the weight matrix.

[0128] S1312. Input the feature matrix into the infectious disease prediction model, and perform an inner product operation on the weight matrix and the feature matrix to generate hidden layer features.

[0129] After obtaining the weight matrix, perform an inner product calculation on the weight matrix and the feature matrix. The process of inner product calculation is to multiply the weight matrix and the feature matrix, and the result of the product is a hidden feature matrix with the same dimension as the weight matrix. This hidden feature matrix is the hidden layer feature. For example, after being encoded by the encoder g and mapped to the hidden layer, the representative feature Z is obtained, and this representative feature Z is the hidden layer feature.

[0130] In some embodiments, after obtaining the hidden layer features, it is necessary to calculate the disease index based on the hidden layer features. Please refer to Figure 7 , Figure 7 which is a schematic flowchart of generating a disease index map for this embodiment. As Figure 7 shown, S1400 includes:

[0131] S1411. Statistically process the data represented by the hidden layer features according to the dimension of the weight matrix to generate a multi-dimensional array;

[0132] After calculating the hidden layer features, each row or each column in the hidden layer features represents a feature of a dimension, and the feature represented by each row of data in the hidden layer features is the same as the dimension of each row in the weight matrix. Therefore, divide the dimension of each row in the hidden layer features according to the data dimension of each row in the weight matrix to generate a multi-dimensional array, and each array represents a row of the matrix corresponding to the hidden layer features.

[0133] S1412. Display the multi-dimensional array in a preset infectious disease index coordinate system to generate a disease index map, where arrays of different dimensions are distinguished and displayed by display colors and / or shapes.

[0134] Map the numbers in each array of the multi-dimensional array through an activation function to map each number into a value between 0 and 1, forming multiple disease indices. Set the disease indices in the corresponding index coordinate system to generate a disease index map.

[0135] The representation methods of the hidden layer features in different dimensions in the disease index map are different. In some embodiments, different dimensions of hidden layer features are represented by different colors. However, the method of distinguishing the display of different dimensions of hidden layer features is not limited to this. According to the adaptability requirements of specific application scenarios, different-shaped coordinate points can be used to represent different dimensions of hidden layer features. In some other embodiments, a combination of colors and shapes is used to distinguish different dimensions of hidden layer features.

[0136] Such as Figure 3 shown, the output results corresponding to the input of the infectious disease prediction model with a time series length n of 150 are shown. It can be seen that the output results of the infectious disease prediction model are relatively evenly distributed in the hidden layer representation space, reflecting the differentiated representation results.

[0137] In some embodiments, in order to more directly register and classify the infectious disease information represented in the disease index map, it is necessary to classify the infection risk of the disease index map through a supervised neural network model. Please refer to Figure 8 , Figure 8 which is a schematic flow chart of the risk classification of the disease index map in this embodiment.

[0138] Such as Figure 8 shown, after S1400, it includes:

[0139] S1421. Input the disease index map into a preset infectious classification model, where the infectious classification model is a neural network model trained to a convergent state through a supervised training method and used for risk classification of infectious diseases;

[0140] Take the disease index map as input and input it into a preset infectious classification model. The infectious classification model is a neural network model trained to a convergent state through a supervised training method and used for risk classification of infectious diseases. The infectious classification model is one of a convolutional neural network model, a deep convolutional neural network model, or a recurrent neural network model, or a variant model of any one of them.

[0141] The risk classification intervals of the infectious classification model include: low risk, medium risk, and high risk. However, the classification intervals of the infectious classification model are not limited to the above classification methods, and the classification intervals can be custom-set according to the needs of specific application scenarios.

[0142] S1422. Read the risk classification result output by the infectious classification model according to the disease index map;

[0143] The data output by the infectious classification model is used to characterize the risk classification result of the disease index map. Therefore, by reading the classification data of the infectious classification model, the risk classification result of the infectious disease corresponding to the disease index map can be obtained.

[0144] S1423. Based on the classification result, match the response strategy corresponding to the classification result in the preset response database.

[0145] According to the risk classification results of different infectious diseases, there are different response strategies for thresholds in the local database of the computer, which are used to respond in a timely manner to infectious diseases with different risk levels. The above response strategies are stored in the response database of the computer, and different response strategies are set with risk level labels. By matching and retrieving through the classification result, the corresponding response strategy can be obtained, and the corresponding response strategy can be executed to issue a warning.

[0146] In some embodiments, for the training method of the infectious disease prediction model, please refer to Figure 9 , Figure 9 which is the schematic diagram of the training process of the infectious disease prediction model in this embodiment. As Figure 9 shown, it includes:

[0147] S2100. Collect a training sample set, where the training sample set includes multiple training samples, and each training sample includes a set of historical infectious disease data;

[0148] Generate training samples through historical data. Specifically, by collecting historical infectious disease data materials, after extracting data with a preset data dimension from the historical infectious disease data, several groups of historical feature data are generated, and each group of historical feature data represents a training sample.

[0149] S2200. Input each of the training samples into the encoder in sequence, so that the encoder performs downsampling processing on the training sample to generate a sample hidden layer feature;

[0150] Input each training sample into the encoder in sequence. After being encoded by the encoder g, it is mapped to the hidden layer to obtain the representation feature Z, where the mapping process is denoted as g(Xm). The processing method of the encoder for each training sample is to perform downsampling, and the representation feature Z obtained for each training sample is the sample hidden layer feature.

[0151] S2300. Input the sample hidden layer feature into the decoder, so that the decoder performs upsampling processing on the sample hidden layer feature to generate sample restored data;

[0152] The decoder f is used to map the characterization feature Z back to the original feature space to obtain a reconstructed sample, and the reconstructed sample is output through the output layer, where the decoding process is denoted as f(Z). The process performed by the decoder is the reverse process of the encoder, and the sample hidden layer features in the low-dimensional space are mapped to the same dimension as the input training sample through upsampling to generate sample restoration data.

[0153] S2400. Calculate the feature distance between the sample restoration data and the historical infectious disease data corresponding to the training sample according to a preset loss function;

[0154] Calculate the feature distance between the sample restoration data and the historical infectious disease data corresponding to the training sample according to the loss function. The characteristics of the loss function are described as follows:

[0155] 。

[0156] The meaning of the objective function is to minimize the reconstruction error of the input monitoring matrix Xm×n, that is, to make the output reconstructed sample as close as possible to the input vector. By minimizing the reconstruction error, the encoder and decoder can be optimized simultaneously, and the relevant parameters can be trained to learn the implicit feature representation Z for the input of the monitoring matrix Xm×n.

[0157] Specifically, in this embodiment, the method for training the parameters of the infectious disease prediction model using the monitoring matrix Xm×n can be the gradient descent method of a deep neural network.

[0158] S2500. Compare the feature distance with a preset distance threshold. When the feature distance is greater than the distance threshold, repeatedly iterate to adjust the weight parameters of the encoder according to the feature distance until the feature distance is less than or equal to the feature threshold.

[0159] Preset a distance threshold, compare the feature distance calculated for each training sample with the preset distance threshold. When the feature distance is greater than the distance threshold, repeatedly iterate to adjust the weight parameters of the encoder according to the feature distance until the feature distance is less than or equal to the feature threshold.

[0160] When the feature distance is less than or equal to the distance threshold, then train the next training sample until the number of training iterations reaches the preset number, and stop training the infectious disease prediction model. At this time, the infectious disease prediction model is trained to convergence.

[0161] For details, please refer to Figure 10 , Figure 10 which is the basic structure schematic diagram of the infectious disease data processing device in this embodiment.

[0162] AsFigure 10 As shown in Figure 10 , an infectious disease data processing device includes: an acquisition module 1100, a processing module 1200, a classification module 1300, an execution module 1400, and an identification module 1500. Among them, the acquisition module 1100 is used to acquire infectious disease data to be processed; the processing module 1200 is used to preprocess the infectious disease data according to preset data dimensions to generate a feature matrix; the classification module 1300 is used to input the feature matrix into a preset infectious disease prediction model, and the infectious disease prediction model performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix; among them, the infectious disease prediction model is pre-trained to a convergence state according to an unlabeled sample model and is used to extract hidden layer information of input data; the execution module 1400 is used to generate a disease index of the infectious disease data according to the hidden layer features; the identification module 1500 is used to write the disease index into a preset Cartesian coordinate system to generate a disease index map of the infectious disease data, and input the disease index map into a preset infectiousness classification model to generate an infectiousness classification result of the infectious disease data.

[0163] The infectious disease data processing device acquires infectious disease data to be processed; preprocesses the infectious disease data according to preset data dimensions to generate a feature matrix; inputs the feature matrix into a preset infectious disease prediction model, and the infectious disease prediction model performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix; among them, the infectious disease prediction model is pre-trained to a convergence state according to an unlabeled sample model and is used to extract hidden layer information of input data; generates a disease index of the infectious disease data according to the hidden layer features. By using the trained infectious disease prediction model for processing and analysis, the calculation of the infectious disease index can be automatically based on regional real data, so it can be applicable to infectious disease indexes with spatial differences; secondly, the downsampling processing process of the infectious disease prediction model is a complex non-linear process, which can more accurately fit the relationship between monitoring matrices of different dimensions, so that the infectious disease index obtained by the model can cover the monitoring data of each dimension as much as possible and adapt to different infectious disease outbreak scenarios. It solves the problem that the accuracy of the infectious disease index obtained by the existing infectious disease index integration method is poor. By visualizing the obtained disease index and performing image recognition on the visualized graph, an infectiousness classification result representing the infectious disease of the infectious disease data is obtained, which can enable the infectiousness classification model to focus on the core indicators of the infectious disease data and make the infectiousness classification result more accurate.

[0164] In some embodiments, the infectious disease data processing device further includes:

[0165] A first acquisition sub-module, configured to acquire a preset acquisition time sequence;

[0166] The first acquisition sub-module is used to sequentially acquire multiple groups of original monitoring data according to the acquisition time sequence within a preset detection duration;

[0167] The first sorting sub-module is used to sort multiple groups of the original monitoring data in ascending order according to the acquisition order to generate infectious disease data.

[0168] In some embodiments, the infectious disease data processing device further includes:

[0169] The first reading sub-module is used to read a preset data dimension;

[0170] The first extraction sub-module is used to sequentially extract data from the infectious disease data according to the data dimension to generate multiple groups of time series vectors corresponding to the infectious disease data;

[0171] The second sorting sub-module is used to sort multiple groups of the time series vectors in sequence according to the corresponding extraction time sequence to generate a feature matrix.

[0172] In some embodiments, the infectious disease data processing device further includes:

[0173] The first calling sub-module is used to call the weight matrix of the infectious disease prediction model, wherein the dimension of the weight matrix is lower than the dimension of the feature matrix;

[0174] The first operation sub-module is used to input the feature matrix into the infectious disease prediction model, so that the weight matrix and the feature matrix perform an inner product operation to generate hidden layer features.

[0175] In some embodiments, the infectious disease data processing device further includes:

[0176] The first statistical sub-module is used to statistically process the data represented by the hidden layer features according to the dimension of the weight matrix to generate a multi-dimensional array;

[0177] The first generation sub-module is used to display the multi-dimensional array in a preset infectious disease index coordinate system to generate a disease index map, wherein arrays of different dimensions are distinguished and displayed by display colors and / or shapes.

[0178] In some embodiments, the infectious disease data processing device further includes:

[0179] The first input sub-module is used to input the disease index map into a preset infectious classification model, wherein the infectious classification model is a neural network model trained to a convergent state by a supervised training method for risk classification of infectious diseases;

[0180] A second reading sub-module, configured to read the risk classification result output by the infectious disease classification model according to the disease index map;

[0181] A first matching sub-module, configured to match a response strategy corresponding to the classification result in a preset response database based on the classification result.

[0182] In some embodiments, the infectious disease data processing device further includes:

[0183] A second acquisition sub-module, configured to acquire a training sample set, where the training sample set includes a plurality of training samples, and each training sample includes a set of historical infectious disease data;

[0184] A second input sub-module, configured to sequentially input each of the training samples into the encoder, so that the encoder performs downsampling processing on the training sample to generate a sample hidden layer feature;

[0185] A third input sub-module, configured to input the sample hidden layer feature into the decoder, so that the decoder performs upsampling processing on the sample hidden layer feature to generate sample restored data;

[0186] A second operation sub-module, configured to calculate a feature distance between the sample restored data and the historical infectious disease data corresponding to the training sample according to a preset loss function;

[0187] A first iteration sub-module, configured to compare the feature distance with a preset distance threshold. When the feature distance is greater than the distance threshold, repeatedly iterate the weight parameters of the encoder according to the feature distance until the feature distance is less than or equal to the feature threshold.

[0188] To solve the above technical problems, an embodiment of the present invention further provides a computer device. Specifically, please refer to Figure 11 , Figure 11 which is the basic structural block diagram of the computer device in this embodiment.

[0189] As Figure 11As shown, it is a schematic diagram of the internal structure of a computer device. The computer device includes a processor, a storage medium, a memory, and a network interface connected through a system bus. Among them, the non-volatile storage medium of the computer device stores an operating system, a database, and computer-readable instructions. The database can store a control information sequence. When the computer-readable instructions are executed by the processor, the processor can implement a method for processing infectious disease data. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. The memory of the computer device can store computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor can execute a method for processing infectious disease data. The network interface of the computer device is used to connect and communicate with a terminal. Those skilled in the art can understand that Figure 11 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0190] In this embodiment, the processor is used to execute Figure 10 the specific functions of the acquisition module 1100, the processing module 1200, the classification module 1300, the execution module 1400, and the recognition module 1500 in. The memory stores the program codes and various types of data required to execute the above modules. The network interface is used for data transmission between the user terminal and the server. The memory in this embodiment stores the program codes and data required to execute all sub-modules in the face image key point detection device. The server can call the program codes and data of the server to execute the functions of all sub-modules.

[0191] A computer device obtains infectious disease data to be processed; preprocesses the infectious disease data according to a preset data dimension to generate a feature matrix; inputs the feature matrix into a preset infectious disease prediction model, and the infectious disease prediction model performs downsampling processing on the feature matrix to generate hidden layer features representing the feature matrix; wherein, the infectious disease prediction model is pre-trained to a convergence state according to an unlabeled sample model and is used to extract hidden layer information of input data; a disease index of the infectious disease data is generated according to the hidden layer features. By using the trained infectious disease prediction model for processing and analysis, the calculation of the infectious disease index can be automatically based on regional real data, so it can be applied to infectious disease indexes with spatial differences; secondly, the downsampling process of the infectious disease prediction model is a complex non-linear process, which can more accurately fit the relationship between monitoring matrices of different dimensions, so that the infectious disease index obtained by the model can cover the monitoring data of each dimension as much as possible and adapt to different infectious disease outbreak scenarios. It solves the problem that the accuracy of the infectious disease index obtained by the existing infectious disease index integration method is poor. By visualizing the obtained disease index and performing image recognition on the visualized graph, a classification result of the infectivity of the infectious disease represented by the infectious disease data is obtained, which can enable the infectivity classification model to focus on the core indicators of the infectious disease data and make the infectivity classification result more accurate.

[0192] The present invention also provides a storage medium storing computer-readable instructions, which when executed by one or more processors, cause the one or more processors to execute the steps of the infectious disease data processing method according to any one of the above embodiments.

[0193] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a Read-Only Memory (ROM), or a Random Access Memory (RAM), etc.

[0194] Those skilled in the art of the present technology can understand that the steps, measures, and solutions in the various operations, methods, and processes discussed in this application can be alternated, changed, combined, or deleted. Further, the other steps, measures, and solutions in the various operations, methods, and processes discussed in this application can also be alternated, changed, rearranged, decomposed, combined, or deleted. Further, the steps, measures, and solutions in the prior art that are the same as those disclosed in this application can also be alternated, changed, rearranged, decomposed, combined, or deleted.

Claims

1. A method for processing infectious disease data, characterized in that, Including: Obtain infectious disease data to be processed; the infectious disease data includes the percentage of infectious disease samples, the number of infectious disease outbreaks, the number of infectious disease cases, and the positive rate of infectious disease viruses; Preprocess the infectious disease data according to preset data dimensions to generate a feature matrix; Invoke the weight matrix of the infectious disease prediction model, where the dimension of the weight matrix is lower than the dimension of the feature matrix; Input the feature matrix into the infectious disease prediction model, and perform an inner product operation on the weight matrix and the feature matrix to generate hidden layer features; among them, the infectious disease prediction model is pre-trained to a convergent state according to an unlabeled sample model and is used to extract the hidden layer information of the input data; Statistically process the data represented by the hidden layer features according to the dimension of the weight matrix to generate a multi-dimensional array; perform weighted summation on each dimension of data in the multi-dimensional array in the hidden layer features to generate disease indices of multiple categories; Write the disease indices of multiple categories generated by the multi-dimensional array into a preset Cartesian coordinate system to generate a disease index graph, where arrays of different dimensions are distinguished by display colors and / or shapes; Input the disease index graph into a preset infectious classification model, where the infectious classification model is a neural network model trained to a convergent state by a supervised training method and is used to classify the risks of infectious diseases; Read the risk classification result output by the infectious classification model according to the disease index graph; Among them, the infectious disease prediction model includes an encoder and a decoder, and the training method of the infectious disease prediction model includes: collecting a training sample set, and making the encoder perform downsampling processing on the training samples to generate sample hidden layer features; inputting the sample hidden layer features into the decoder, and making the decoder perform upsampling processing on the sample hidden layer features to generate sample restoration data.

2. The infectious disease data processing method according to claim 1, wherein The method for obtaining the infectious disease data to be processed includes: Obtain a preset acquisition time sequence; Collect multiple groups of original monitoring data in sequence according to the acquisition time sequence within a preset detection duration; Arrange the multiple groups of original monitoring data in ascending order according to the acquisition order to generate infectious disease data.

3. The infectious disease data processing method according to claim 2, wherein The method for preprocessing the infectious disease data according to preset data dimensions to generate a feature matrix includes: Read the preset data dimensions; Extract data from the infectious disease data in sequence according to the data dimensions to generate multiple groups of time sequence vectors corresponding to the infectious disease data; Sort the multiple groups of time sequence vectors in sequence according to the corresponding extraction time sequence to generate a feature matrix.

4. The infectious disease data processing method according to claim 1, wherein The infectious disease data processing method further includes: Match a response strategy corresponding to the classification result in a preset response database based on the classification result.

5. An infectious disease data processing device, characterized in that, Including: An acquisition module, configured to obtain infectious disease data to be processed; the infectious disease data includes the percentage of infectious disease samples, the number of infectious disease outbreaks, the number of infectious disease cases, and the positive rate of infectious disease viruses; A processing module, configured to preprocess the infectious disease data according to preset data dimensions to generate a feature matrix; A classification module, configured to call the weight matrix of the infectious disease prediction model, wherein the dimension of the weight matrix is lower than that of the feature matrix; input the feature matrix into the infectious disease prediction model to perform an inner product operation on the weight matrix and the feature matrix to generate hidden layer features; wherein, the infectious disease prediction model is pre-trained to a converged state according to an unlabeled sample model and is used to extract the hidden layer information of the input data; An execution module, configured to statistically process the data represented by the hidden layer features according to the dimension of the weight matrix to generate a multi-dimensional array; perform weighted summation on each dimension data in the multi-dimensional array in the hidden layer features to generate disease indices of multiple categories; An identification module, configured to write the disease indices of multiple categories generated by the multi-dimensional array into a preset Cartesian coordinate system to generate a disease index map, wherein arrays of different dimensions are distinguished and displayed by display colors and / or shapes; Input the disease index map into a preset infectiousness classification model, wherein the infectiousness classification model is a neural network model trained to a converged state by a supervised training method and is used to perform risk classification on infectious diseases; Read the risk classification result output by the infectiousness classification model according to the disease index map; Wherein, the infectious disease prediction model includes an encoder and a decoder, and the classification module is specifically configured to: collect a training sample set to enable the encoder to perform downsampling processing on the training samples to generate sample hidden layer features; input the sample hidden layer features into the decoder to enable the decoder to perform upsampling processing on the sample hidden layer features to generate sample restored data.

6. A computer device, including a memory and a processor, wherein computer-readable instructions are stored in the memory, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the infectious disease data processing method according to any one of claims 1 to 4.

7. A storage medium storing computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the infectious disease data processing method according to any one of claims 1 to 4.

Images