Method, device, electronic equipment and medium for predicting the number of infectious diseases

Abstract

The disclosed embodiment provides a method, device, electronic device and computer-readable storage medium for predicting the number of infectious diseases, belonging to the field of medical data technology. The method for predicting the number of infectious diseases includes: correcting the original medical data related to the target disease, obtaining the corrected data corresponding to the original medical data; determining multidimensional feature data in combination with the corrected data and the associated data corresponding to the target disease; predicting according to the multidimensional feature data corresponding to the reference time before the target time, so as to determine the number of cases belonging to the target disease for the target time. The disclosed embodiment can improve the accuracy of case number prediction.

Description

[0001] Description of the case

[0002] This application is a divisional application of a patent application with an application date of March 19, 2020, application number 202010194757.4, and name “Infectious Disease Number Prediction Method, Device, Electronic Equipment and Medium”. Background Art

[0004] After an infectious disease occurs in a country or region, accurately predicting the number of infections and future trends is extremely important for controlling the disease.

[0005] In the related art, fixed parameters are generally used to build models and predict the number of cases. In this way, fixed parameters are difficult to describe complex situations, so there are certain limitations. In addition, the related art can only use historical data for prediction, which has a certain lag, so it may lead to missed identification, resulting in poor accuracy and reliability of the prediction results.

[0006] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0007] The purpose of the embodiments of the present disclosure is to provide a method for predicting the number of infectious diseases, an apparatus for predicting the number of infectious diseases, an electronic device and a computer-readable storage medium, thereby overcoming the problems of low accuracy and poor reliability of data prediction at least to a certain extent.

[0008] Other features and advantages of the embodiments of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present invention.

[0009] According to one aspect of an embodiment of the present disclosure, a method for predicting the number of people with an infectious disease is provided, comprising: correcting original medical data related to a target disease to obtain corrected data corresponding to the original medical data; determining multidimensional feature data in combination with the corrected data and associated data corresponding to the target disease; and predicting based on the multidimensional feature data corresponding to a reference duration before a target time to determine the number of cases belonging to the target disease at the target time.

[0010] In an exemplary embodiment of the present disclosure, the original medical data related to the target disease is corrected to obtain the corrected data corresponding to the original medical data, including: based on the exponential averaging technology from back to front in time, logical operations are performed on the original medical data of the unit time to be processed and the corrected data of the next unit time adjacent to the unit time to be processed, so as to obtain the corrected data of the unit time to be processed.

[0011] In an exemplary embodiment of the present disclosure, the correcting of the original medical data related to the target disease to obtain the corrected data corresponding to the original medical data includes: averaging the increments of the original medical data per unit time to obtain the corrected data corresponding to the original medical data.

[0012] In an exemplary embodiment of the present disclosure, the prediction is performed based on the multidimensional feature data corresponding to the reference time before the target time to determine the number of cases belonging to the target disease for the target time, including: inputting the multidimensional feature data corresponding to the reference time into a trained machine learning model for prediction to determine the number of cases at the target time.

[0013] In an exemplary embodiment of the present disclosure, the method further includes: training the machine learning model through historical data and actual results of the historical data to obtain the trained machine learning model.

[0014] In an exemplary embodiment of the present disclosure, the machine learning model is trained through historical data and the actual results of the historical data to obtain the trained machine learning model, including: inputting historical data into the machine learning model for prediction to obtain the prediction results of the historical data; adjusting the weight parameters of the machine learning model according to the comparison results of the prediction results and the actual results until the prediction results are consistent with the actual results to obtain the trained machine learning model.

[0015] In an exemplary embodiment of the present disclosure, the historical data is input into a machine learning model for prediction to obtain a prediction result of the historical data, including: correcting the historical data related to the target disease to obtain corrected data corresponding to the historical data; determining reference multidimensional feature data in combination with the corrected data corresponding to the historical data and the historical associated feature data corresponding to the target disease; and performing quantitative prediction on the reference multidimensional corrected data through the machine learning model to obtain a prediction result of the historical data.

[0016] According to one aspect of an embodiment of the present disclosure, there is provided a device for predicting the number of people with an infectious disease, comprising: a data correction module, for correcting original medical data related to a target disease, and obtaining corrected data corresponding to the original medical data; a feature extraction module, for determining multidimensional feature data in combination with the corrected data and associated data corresponding to the target disease; and a quantity prediction module, for predicting based on the multidimensional feature data corresponding to a reference duration before a target time, so as to determine the number of cases belonging to the target disease for the target time.

[0017] According to one aspect of an embodiment of the present disclosure, there is provided a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for predicting the number of infectious diseases as described in any one of the above is implemented.

[0018] According to one aspect of an embodiment of the present disclosure, there is provided an electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any one of the above-mentioned methods for predicting the number of infectious diseases by executing the executable instructions.

[0019] In the infectious disease number prediction method, device, electronic device and computer-readable storage medium provided by the embodiments of the present disclosure, the original medical data is corrected to obtain the corrected data corresponding to the original medical data, and then the multidimensional feature data used to represent the characteristics of the original medical data is obtained based on the corrected data, and the number prediction is performed based on the multidimensional feature data of the corrected data within the reference time to determine the number of cases within the target time. On the one hand, since the original medical data can be corrected, the data deviation caused by lag is reduced and the accuracy of the data is improved. On the other hand, the number of cases can be predicted by the multidimensional feature data corresponding to the reference time, and the case data can be predicted from multiple dimensions, avoiding the limitations of the related technology that can only be predicted based on fixed parameters, and can comprehensively consider the characteristics of multiple dimensions, reduce errors, and then accurately predict the number of cases corresponding to the target time, and can achieve dynamic prediction and improve the reliability of the prediction results.

[0020] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings herein are incorporated into and constitute a part of the specification, showing embodiments consistent with the present invention, and together with the specification, are used to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0022] Figure 1 A system architecture diagram for executing a method for predicting the number of people with an infectious disease according to an embodiment of the present disclosure is schematically shown;

[0023] Figure 2 The flowchart of the infectious disease number prediction method according to the embodiment of the present disclosure is schematically shown;

[0024] Figure 3The following is a schematic diagram showing the process flow of model training in an embodiment of the present disclosure;

[0025] Figure 4 A schematic diagram schematically shows the number of cases for determining the target time in an embodiment of the present disclosure;

[0026] Figure 5 A block diagram of a device for predicting the number of people with an infectious disease according to an embodiment of the present disclosure is schematically shown;

[0027] Figure 6 The block diagram of an electronic device used to implement the above-mentioned infectious disease number prediction method is schematically shown. DETAILED DESCRIPTION

[0028] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the present invention will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0029] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present invention. However, those skilled in the art will appreciate that the technical solution of the present invention can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring various aspects of the present invention.

[0030] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0031] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0032] Figure 1 The figure schematically shows a system architecture diagram for executing the infectious disease number prediction method in the embodiment of the present disclosure.

[0033] like Figure 1As shown, the system architecture 100 may include a first end 101, a network 102, and a second end 103. Among them, the first end 101 may be a client, for example, it may be various handheld devices (smart phones) with display functions, desktop computers, vehicle-mounted devices, etc. The network 102 is used to provide a medium for a communication link between the first end 101 and the second end 103. The network 102 may include various connection types, such as a wired communication link, a wireless communication link, etc. In the embodiment of the present disclosure, the network 102 between the first end 101 and the second end 103 may be a wired communication link, for example, a communication link may be provided through a serial port connection line, or a wireless communication link, and a communication link may be provided through a wireless network. The second end 103 may be a client, such as a portable computer, a desktop computer, a smart phone, and other terminal devices with data processing functions, for predicting the number of cases of the target disease. Among them, when the first end 101 and the second end 103 are both clients, they may be the same client. The second end 103 may also be a server, which is not limited here.

[0034] It should be understood that Figure 1 The number of the first end, the network and the second end in the embodiment is only for illustration. According to the implementation requirements, there can be any number of clients, networks and servers.

[0035] It should be noted that the infectious disease number prediction method provided in the embodiment of the present disclosure can be completely executed by the second end 103 , and accordingly, the infectious disease number prediction device can be set in the second end 103 .

[0036] Based on the above system architecture, the present disclosure provides a method for predicting the number of people with infectious diseases. The method can be applied to the application scenario of processing medical data to predict the data. The execution subject of the method can be a server or a terminal with computing power. Figure 2 As shown, the infectious disease population prediction method includes steps S210 to S230, which are described in detail as follows:

[0037] In step S210, the original medical data related to the target disease is corrected to obtain corrected data corresponding to the original medical data;

[0038] In step S220, multi-dimensional feature data is determined by combining the correction data and the associated data corresponding to the target disease;

[0039] In step S230, prediction is performed based on the multidimensional feature data corresponding to the reference time before the target time to determine the number of cases belonging to the target disease at the target time.

[0040] In the infectious disease number prediction method provided by the embodiment of the present disclosure, on the one hand, since the original medical data can be corrected, the data deviation caused by lag is reduced and the accuracy of the data is improved. On the other hand, the number of cases can be predicted through multi-dimensional feature data, and the case data can be predicted from multiple dimensions, avoiding the limitation caused by the prediction based on fixed parameters in the related technology, and comprehensively considering the characteristics of multiple dimensions, reducing errors, and accurately predicting the number of cases corresponding to the target time, which can realize dynamic prediction and improve the reliability of the prediction results.

[0041] Next, the infectious disease number prediction method in the embodiment of the present disclosure is explained in detail with reference to the accompanying drawings.

[0042] In step S210, the original medical data related to the target disease is corrected to obtain corrected data corresponding to the original medical data.

[0043] In the disclosed embodiments, the target disease may be a contagious disease (infectious disease), for example, various epidemics or various infectious influenza, infectious diseases, and the like. The target disease may be for a certain region or for all regions, which is not limited here. Original medical data refers to the data about the target disease that has been published before the current time, and the original medical data may be historical confirmed data, that is, the number of cases belonging to the target disease. In order to process the data, the original medical data related to the target disease may be obtained from the database of an institution such as a hospital or a medical institution and a medical center that can diagnose the target disease, or the original medical data may be obtained from the Internet. In order to avoid the problem that the number of confirmed cases is less than the actual number due to the lag in the initial medical response to the discovery of the target disease, the original medical data related to the target disease may be corrected in the disclosed embodiments to make the original medical data more consistent with the actual data.

[0044] In the disclosed embodiment, the original medical data can be corrected by smoothing the original medical data. Specifically, the method for correcting the original medical data related to the target disease may include: based on the exponential averaging technology from back to front in time, logical operations are performed on the original medical data of the unit time to be processed and the corrected data of the next unit time adjacent to the unit time to be processed, and the corrected data of the unit time to be processed is obtained. Among them, the unit of time can be a day, so the unit time to be processed can be any day before the target time, for example, it can be every day. The next unit time is used to represent the next day of every day. It should be noted that the unit time to be processed is the time when the original medical data has been determined, and the unit time to be processed can be the time when the original medical data has been determined on the second day and any day thereafter. Exponential averaging technology is a technology often used in gradient descent algorithms. Exponential averaging realizes data smoothing by calculating the dynamic average of historical values ​​and current values, so that noise data can be well suppressed and the algorithm can be more stable.

[0045] Specifically, we can assume that the original medical data is [y1,y2,...,y i ,...,y t ], the corrected data is [n1,n2,...,n i ,...,n t ]. In this case, the corrected data corresponding to the original medical data can be determined by the following method and formula (1), where:

[0046] Let y t =n t , taking i as input, starting from i=t-1, and ending the loop when i=1.

[0047] n i =β*n i+1 +(1-β)*y i Formula (1)

[0048] Here, β is a constant less than 1.

[0049] In addition, other smoothing techniques can also be used to correct the original medical data. Specifically, the increment of the original medical data per unit time can be averaged to obtain the corrected data corresponding to the original medical data. The increment refers to the increase in case data per unit time relative to the previous unit time. However, due to the lag in the initial stage, the increment of each unit time may not be generated by this unit time, but by multiple unit times. Therefore, the increase in the original medical data per unit time can be distributed to all unit times. All unit times here refer to the unit time for which the original medical data has been determined, that is, all days for which the number of cases has been determined. When distributing, it can be distributed evenly or proportionally, and it can be processed according to various logical operations when distributing, and there is no special limitation here.

[0050] For example, if the target disease starts on March 1 and the increment on March 3 is 900, the 900 increments can be evenly distributed to all days including March 3, and the revised data for each unit time can be the sum of the original medical data and the increments allocated for each unit time.

[0051] In the disclosed embodiment, by correcting the original medical data of each unit time according to the data of the later unit time, it is possible to avoid ignoring the problem that the number of confirmed cases is far lower than the actual number due to the lag in medical response in the early stage of the outbreak, and to correct the deviation in the case data, making the corrected data more accurate, thereby making the learned and trained model more accurate and more in line with the actual situation.

[0052] Continue to refer Figure 2 As shown in , in step S220, multidimensional feature data is determined by combining the correction data and the associated data corresponding to the target disease.

[0053] In the embodiment of the present disclosure, after obtaining the correction data, feature data for representing the development state of the target disease can be obtained on the basis of the correction data, and the feature data here can be used to uniquely represent the characteristics corresponding to the target disease. In order to avoid the problem of limitations caused by predicting from fixed parameters in the related art, in the embodiment of the present disclosure, the multidimensional feature data can be determined by combining the correction data that conforms to the actual situation and the associated data corresponding to the target disease. The associated data corresponding to the target disease can be data related to the development state of the target disease (such as clinical data, treatment data, and input data, etc.), which is specifically used to assist in predicting the target disease. The associated data can be derived from data of multiple dimensions so as to provide comprehensive auxiliary effects. The multiple dimensions of the associated data can be at least one or more of the treatment dimension, the user dimension, the input dimension, and the control intensity. The treatment dimension can be the number of hospital beds, the user dimension can be the number of days from onset to diagnosis (i.e., the incubation period), the input dimension can be the number of inflows of foreign population, and the control intensity dimension can be the activity of people flow, etc. The associated data can be dynamic data, that is, the associated data of each day is updated in real time. It should be noted that the associated data can also be derived from more dimensions, which can be determined according to actual needs. The multidimensional feature data refers to the multidimensional feature data of each unit time, and the multidimensional feature data of each unit time may be different, which is specifically determined according to the actual correction data and the associated data.

[0054] In the disclosed embodiment, the associated data is derived from the number of hospital beds, the number of days from onset to diagnosis (i.e., the incubation period), and the number of foreign population inflows as an example for explanation. The associated data can be obtained based on the statistical information in the database. Specifically, after obtaining the corrected data and the associated data, the corrected data and the associated data can be spliced ​​and combined. When combining the corrected data and the associated data, they can be spliced ​​in any order, as long as they can include all the corrected data and the associated data, and no specific limitation is made here.

[0055] For example, the multidimensional feature data obtained by combining the correction data and the associated data can be expressed as x i , and the multidimensional feature data x i =[n i ,m i ,b i ,d i ]. Among them, n i It is used to indicate the number of historical infections. The number of historical infections here refers to the above-mentioned revised data. m i Used to indicate the influx of foreign population, b i Used to indicate the number of hospital beds, d i Used to indicate the number of days from onset to diagnosis.

[0056] In the disclosed embodiments, multi-dimensional feature data for representing the development status of the target disease is obtained by combining the corrected data with the associated data corresponding to the target disease, thereby avoiding the problem of inaccurate prediction results caused by only being able to make predictions based on one dimension of the historical number of infections in the related art. The function of comprehensive prediction from multiple dimensions can be realized, and complex situations can be dynamically described through multiple dimensions, thereby improving accuracy and usability.

[0057] Continue to refer Figure 2 As shown in , in step S230, prediction is performed based on the multidimensional feature data corresponding to the reference time before the target time to determine the number of cases belonging to the target disease for the target time.

[0058] In the disclosed embodiment, the reference duration refers to the duration before the target time, and the unit time composition of the original medical data and the revised data has been determined. The reference duration may include one unit time or multiple unit times, and the reference duration includes several unit times, which can be specifically determined according to the position of the target time in the development process of the entire target disease, and the position can be described by the number of unit times. The target time refers to the unit time (i.e., the next day) at the current moment where the number is not determined. Specifically, if the target time is earlier than the 4th day, all unit times before the 4th day can be used as the reference duration. If the target time is later than the 4th day (for example, the 5th day), the preset number of unit times before the 5th day can be used as the reference duration. Among them, the preset number of unit times can be 3 days or 4 days, etc. After testing, it is shown that when the preset number of unit times is 4 days, the result obtained is more accurate. On this basis, if the target time is the 2nd day, the reference duration can be determined based on a unit time before the 2nd day; if the target time is the 3rd day, the reference duration is the previous 2 days; if the target time is the 6th day, the reference duration is the 4 days before and adjacent to the 6th day. In the disclosed embodiment, the reference duration is determined by the location of the target time, which can make the determined data more valuable for reference, thereby making the result more accurate.

[0059] After determining the reference duration, the multidimensional feature data corresponding to each unit time contained in the reference duration can be obtained, and the multidimensional feature data of the unit time of the reference duration can be combined to obtain a multidimensional feature data, and data prediction can be performed based on the multidimensional feature data. For example, the multidimensional feature data of the first four days adjacent to the i-th day [x i-4 ,x i-3 ,x i-2 ,x i-1 ] to predict the number of cases n at the target time (i-th day) i. It should be noted that if the multidimensional feature data of the previous 4 days are not determined, it is necessary to first estimate the multidimensional feature data of the previous 4 days according to the above method, and further estimate the number of cases at the target time (the i-th day) based on the estimated multidimensional feature data of the previous 4 days. For example, if only 3 days of data are determined, the number of cases on the 4th day can be estimated based on the multidimensional feature data of the previous 3 days. Further, if the i-th day is the 5th day, the number of cases at the target time (the 5th day) can be estimated based on the multidimensional feature data of the 4 days before the 5th day.

[0060] When predicting the number of cases at the target time, the multidimensional feature data corresponding to the reference duration can be input into the trained machine learning model for prediction to determine the number of cases at the target time. The number of cases refers to the number of infectious diseases. The machine learning model here can be any suitable model, such as a long short-term memory network or a tree model, etc. The long short-term memory network is used as an example for illustration in the disclosed embodiment.

[0061] Before using a trained machine learning model to predict multidimensional feature data, in order to improve the accuracy of data prediction, the machine learning model can be first trained to obtain a trained machine learning model.

[0062] Figure 3 The flowchart of model training is shown schematically in Figure 3 As shown in , it may specifically include step S310 and step S320, wherein:

[0063] In step S310, the historical data is input into a machine learning model for prediction to obtain a prediction result of the historical data;

[0064] In step S320, the weight parameters of the machine learning model are adjusted according to the comparison result between the predicted result and the actual result until the predicted result is consistent with the actual result, so as to obtain the trained machine learning model.

[0065] In the disclosed embodiment, historical data can be used as input to a machine learning model, and the prediction result of the historical data can be determined by the machine learning model. The prediction result here can be the number of cases corresponding to the historical data. In order to ensure the accuracy of the data, the historical data can be corrected to obtain the corrected data corresponding to the historical data, and then the machine learning model can be trained on the basis of the corrected data. Specifically, predicting the historical data can include the following steps: correcting the historical data related to the target disease to obtain the corrected data corresponding to the historical data; determining the reference multidimensional feature data by combining the corrected data corresponding to the historical data and the historical associated feature data corresponding to the target disease; and predicting the reference multidimensional corrected data by the machine learning model to obtain the prediction result of the historical data. Among them, the historical data can be the original medical data of all unit times before the target time. The historical data can be corrected by the same correction method as the original medical data. Specifically, based on the exponential average technology from back to front in time, the original medical data and the corrected data of the next unit time adjacent to the unit time to be processed can be logically operated to obtain the corrected data of the unit time to be processed. Or the increment of the original medical data in the unit time is averaged to obtain the corrected data corresponding to the original medical data contained in the historical data.

[0066] After obtaining the correction data of the historical data, the correction data of the historical data and the reference associated data of the target disease per unit time to which the historical data belongs can be combined to obtain reference multidimensional feature data. The reference associated data here can be partially identical to the associated data, and the reference multidimensional feature data can also be partially identical to the multidimensional feature data.

[0067] Furthermore, the reference multidimensional feature data can be input into a machine learning model to predict the number of cases per unit time to which the historical data belongs through processing processes such as convolution operations.

[0068] The parameters of the machine learning model can be further adjusted and trained according to the actual results and predicted results of the historical data until the predicted results of the historical data obtained by the machine learning model are consistent with the actual results, and the parameters when the results are consistent are used as the final parameters of the machine learning model to form a trained machine learning model. Historical data refers to the corrected data of all unit times before the target time. The actual result refers to the actual number of unit times to which the historical data belongs.

[0069] After obtaining the trained machine learning model, the multidimensional feature data corresponding to the test time before the target time can be input into the trained machine learning model for prediction to determine the number of cases at the target time. Next, the machine learning model is taken as a long short-term memory network as an example for explanation.

[0070] The long short-term memory network in the embodiment of the present disclosure may include multiple gates, which may specifically include but are not limited to a forget gate, an update gate, and an output gate. Among them, the forget gate is used to determine the discarded information. The forget gate reads the value passed from the last weight block and the value input this time, and outputs a number between 0 and 1, 1 means completely retained, and 0 means completely discarded. This function can help to dynamically determine whether to remember the value of the last input and whether to consider the interaction between the values ​​of the two inputs.

[0071] The update gate, also known as the input gate, is used to determine how much new information to add to the model. This process requires two specific steps: first, the sigmoid function layer needs to decide which input information needs to be updated; a tanh layer generates a vector as an alternative, which is used for update content. Next, these two parts are combined to update the state of the model.

[0072] The output gate is used to determine what value to output. This output is based on the entire model. First, a sigmoid function layer is used to determine which part of the model state will be output. Then the model state is processed through tanh to obtain a value between -1 and 1. This value is multiplied by the output of the sigmoid gate, and finally only the determined output part will be output.

[0073] For long short-term memory networks, it can also include cell states. At time t, input multidimensional feature data x t and the hidden state h at time t-1 t-1 , under the joint action of weight W, weight U and bias vector b, the state of the update gate, output gate and forget gate at time t is generated, which can be expressed by u t , o t and f t Further, the cell state c at time t is generated under the action of the cell state at time t-1 t Finally, at time t, the cell state c t AND output gate o t Under the action of, the hidden state h at time t is generated t , which in turn affects the internal changes of the neurons in the long short-term memory network at time t+1. The specific process can be referred to as shown in formula (2):

[0074]

[0075] Among them, and represent the weight of the unit state, refers to the multidimensional feature data at time t, refers to the hidden state at time t-1, represents the bias vector of the unit state, and represent the weight of the forget gate, represents the bias vector of the forget gate, and represent the weight of the update gate, represents the bias vector of the update gate, and represent the weight of the output gate, represents the bias vector of the output gate, represents the unit state at time t-1, and represents the unit state at time t, represents the forget gate at time t, represents the update gate at time t, represents the output gate at time t, represents the hidden state at time t, and represents the predicted number of people (i.e., the predicted number of cases).

[0076] In the embodiment of the present disclosure, based on the long short-term memory network, the multidimensional feature data of multiple unit times in the reference time are linearly transformed to calculate the number of cases in the target time through the trained long short-term memory network. For example, the reference time is the first 4 days of the target time. The multidimensional feature data of the first 4 days can be represented as x i-4 , x i-3 , x i-2 , x i-1 . refer to Figure 4 The prediction process diagram shown in the figure is to transform the multidimensional feature data x on the first day into i-4 Through the hidden state h i-4 Perform nonlinear processing (nonlinear transformation) to transform the multidimensional feature data x of the second day i-3 Through the hidden state h i-3 Perform nonlinear processing and transform the multidimensional feature data x on the third day i-2 Through the hidden state h i-2 Perform nonlinear processing until the multidimensional feature data x on the 4th day is i-1 Through the hidden state h i-1 Perform nonlinear processing to obtain the output of the long short-term memory network Until then, the output of the long short-term memory network is used as the number of cases of the target disease corresponding to the target time. Among them, the hidden state is updated in real time, so the hidden state corresponding to each unit time may be different.

[0077] Figure 4 The technical solution in the invention makes predictions by referring to the multi-dimensional feature data of multiple time units contained in the duration and the hidden state of each time unit, and can combine historical data to make quantity predictions from multiple dimensions, so as to make dynamic predictions more accurately. Moreover, the method in the embodiment of the present disclosure can be applied to various regions (such as a city, a province, or a country, etc.), which increases the scope of application and improves universality.

[0078] In the disclosed embodiment, when using a long short-term memory network that can be used for time series analysis to predict the number of cases at the target time, predictions can be made based on the multidimensional feature data of multiple unit times contained in the reference time. Since the multidimensional feature data includes data of multiple dimensions, the parameters of the model can be dynamically learned based on the data to obtain a well-trained machine learning model with better performance, and then accurate predictions can be made based on the model. Therefore, the number of cases at the target time can be predicted comprehensively and accurately, and the number of predicted cases obtained is more accurate and more in line with the actual situation, and will not be affected by the lag of the data, so the reliability of the prediction results can be improved. Prediction through features of multiple dimensions can describe more complex situations, realize dynamic predictions that follow the flow of people and changes in protective measures, and have a wider range of applications. The LSTM (Long Short-Term Memory, long short-term memory network) method can automatically learn the parameters of the model through historical data, can depict more complex situations in the real world, is more practical, avoids limitations, and increases usability.

[0079] In the embodiment of the present disclosure, a device for predicting the number of people with infectious diseases is also provided. Figure 5 As shown in , the infectious disease number prediction device 500 may include the following modules:

[0080] The data correction module 501 is used to correct the original medical data related to the target disease and obtain the corrected data corresponding to the original medical data;

[0081] A feature extraction module 502 is used to determine multi-dimensional feature data by combining the correction data and the associated data corresponding to the target disease;

[0082] The quantity prediction module 503 is used to make a prediction based on the multi-dimensional feature data corresponding to the reference time before the target time to determine the number of cases belonging to the target disease at the target time.

[0083] In an exemplary embodiment of the present disclosure, the data correction module includes: a first correction module, which is used to perform logical operations on the original medical data of the unit time to be processed and the corrected data of the next unit time adjacent to the unit time to be processed based on the exponential averaging technology from back to front in time, so as to obtain the corrected data of the unit time to be processed.

[0084] In an exemplary embodiment of the present disclosure, the data correction module includes: a second correction module, which is used to average the increments of the original medical data per unit time to obtain corrected data corresponding to the original medical data.

[0085] In an exemplary embodiment of the present disclosure, the quantity prediction module includes: a data prediction module, which is used to input the multidimensional feature data corresponding to the reference duration into a trained machine learning model for prediction, so as to determine the number of cases at the target time.

[0086] In an exemplary embodiment of the present disclosure, the device also includes: a model training module, which is used to train the machine learning model through historical data and real results of the historical data to obtain the trained machine learning model.

[0087] In an exemplary embodiment of the present disclosure, the model training module includes: a historical data prediction module, which is used to input historical data into a machine learning model for prediction to obtain a prediction result of the historical data; a parameter adjustment module, which is used to adjust the weight parameters of the machine learning model according to the comparison result between the prediction result and the actual result until the prediction result is consistent with the actual result to obtain the trained machine learning model.

[0088] In an exemplary embodiment of the present disclosure, the historical data prediction module is configured to: correct the historical data related to the target disease and obtain the corrected data corresponding to the historical data; determine the reference multidimensional feature data by combining the corrected data corresponding to the historical data and the historical associated feature data corresponding to the target disease; and perform quantitative prediction on the reference multidimensional corrected data through the machine learning model to obtain the prediction result of the historical data.

[0089] It should be noted that the various functional modules of the infectious disease number prediction device of the embodiment of the present disclosure are the same as the steps of the example embodiment of the infectious disease number prediction method mentioned above, and therefore will not be repeated here.

[0090] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0091] In addition, although the steps of the method of the present invention are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0092] In an exemplary embodiment of the present invention, an electronic device capable of implementing the above method is also provided.

[0093] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0094] Refer to the following Figure 6 The electronic device 600 according to this embodiment of the present invention is described. Figure 6 The electronic device 600 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0095] like Figure 6 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including the storage unit 620 and the processing unit 610), and a display unit 640.

[0096] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 610 can perform the following steps: Figure 2 Follow the steps shown in .

[0097] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0098] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0099] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0100] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 660. As shown, the network adapter 660 communicates with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0101] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation method of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including a number of instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the infectious disease number prediction method according to the implementation method of the present invention.

[0102] In an exemplary embodiment of the present invention, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of this specification is stored. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to perform the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of this specification.

[0103] The program product for implementing the above method according to an embodiment of the present invention may adopt a portable compact disk read-only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus, or device.

[0104] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0105] Computer readable signal media may include a data signal propagated in baseband or as part of a carrier wave, wherein a readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0106] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0107] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0108] Other embodiments of the invention will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

Claims

1. A method for predicting the number of people infected with an infectious disease, characterized in that: include: Distribute the increase of the original medical data per unit time related to the target disease to all unit times, and obtain the revised data per unit time corresponding to the original medical data, wherein the revised data per unit time is the sum of the original medical data and the increment distributed to each unit time; Combining the correction data and the associated data corresponding to the target disease to obtain multi-dimensional feature data; The associated data corresponding to the target disease is data related to the development status of the target disease, and the associated data includes at least one or more of a treatment dimension, a user dimension, an input dimension, and a control intensity; Obtain multidimensional feature data corresponding to each adjacent unit time in the reference time before the target time; Combining the multidimensional feature data per unit time of the reference duration to obtain multidimensional feature data corresponding to the reference duration; The multidimensional feature data corresponding to the reference duration is input into a trained machine learning model for prediction to determine the number of cases at the target time.

2. The method for predicting the number of infectious diseases according to claim 1, characterized in that: The method further comprises: Based on the exponential averaging technology from back to front in time, logical operations are performed on the original medical data of the unit time to be processed and the revised data of the next unit time adjacent to the unit time to be processed to obtain the revised data of the unit time to be processed.

3. The method for predicting the number of infectious diseases according to claim 1, characterized in that: The method further comprises: The machine learning model is trained using historical data and actual results of the historical data to obtain the trained machine learning model.

4. The method for predicting the number of infectious diseases according to claim 3, characterized in that: The training of the machine learning model by using the historical data and the real results of the historical data to obtain the trained machine learning model includes: Inputting historical data into a machine learning model for prediction to obtain prediction results of the historical data; The weight parameters of the machine learning model are adjusted according to the comparison result between the predicted result and the actual result until the predicted result is consistent with the actual result, so as to obtain the trained machine learning model.

5. The method for predicting the number of infectious diseases according to claim 4, characterized in that: The inputting of historical data into the machine learning model for prediction to obtain the prediction result of the historical data includes: Correcting historical data related to the target disease and obtaining corrected data corresponding to the historical data; Determine reference multidimensional feature data by combining the correction data corresponding to the historical data and the historical associated feature data corresponding to the target disease; The reference multidimensional feature data is quantitatively predicted by the machine learning model to obtain a prediction result of the historical data.

6. A device for predicting the number of people infected with an infectious disease, characterized in that: include: A data correction module, used to distribute the increase of the original medical data per unit time related to the target disease to all unit times, and obtain the correction data per unit time corresponding to the original medical data, wherein the correction data per unit time is the sum of the original medical data and the increment allocated for each unit time; A feature extraction module, used to combine the correction data and the associated data corresponding to the target disease to obtain multidimensional feature data, and obtain the multidimensional feature data corresponding to each adjacent unit time in the reference time before the target time, and combine the multidimensional feature data of the unit time of the reference time to obtain the multidimensional feature data corresponding to the reference time; The associated data corresponding to the target disease is data related to the development status of the target disease, and the associated data includes at least one or more of a treatment dimension, a user dimension, an input dimension, and a control intensity; The quantity prediction module is used to input the multidimensional feature data corresponding to the reference duration into a trained machine learning model for prediction to determine the number of cases at the target time.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the number of infectious diseases as described in any one of claims 1 to 5 is implemented.

8. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; Wherein, the processor is configured to execute the infectious disease number prediction method according to any one of claims 1-5 by executing the executable instructions.

Images