Disease Monitoring and Early Warning Method and Device, Terminal Device, Storage Medium

Abstract

The present disclosure relates to the technical field of data analysis, and particularly relates to a disease monitoring and early warning method, a storage medium, and a terminal device. The method includes: obtaining feature data for a preset monitoring period; determining a single feature sequence set according to the feature data in combination with feature weight data corresponding to a disease to be evaluated, wherein the single feature sequence includes the weighted sum result of each feature and the corresponding feature weight for each day within the monitoring period; inputting the single feature sequence set into a trained risk assessment model to obtain an outbreak risk assessment result of the disease to be evaluated within the preset monitoring period. This solution can improve the accuracy of the outbreak risk early warning for infectious diseases.

Description

Background Art

[0002] The spread of infectious diseases can have a significant impact on people's health, lives, and the economy of society. This makes the prevention and early warning of infectious diseases an important research topic. In existing technical solutions, algorithms such as the moving percentile method and the cumulative sum control chart method are used to establish an infectious disease model to achieve the purpose of early warning. However, such algorithms require some indicators to reach certain early warning indicators, or the data to show a certain trend before they will give an alarm; it also indicates that the infectious disease has reached a certain scale of infection at this time. Therefore, there is a certain lag in such an early warning method.

[0003] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0004] The purpose of the present disclosure is to provide a disease monitoring and early warning method, device, storage medium, and terminal device, so as to at least to some extent overcome the limitations and defects of related technologies and be able to accurately evaluate the spread risk of diseases.

[0005] Other features and advantages of the present disclosure will become apparent through the following detailed description, or will be learned in part through the practice of the present disclosure.

[0006] According to a first aspect of the present disclosure, a disease monitoring and early warning method is provided, including:

[0007] Obtain the characteristic data of a preset monitoring period;

[0008] According to the characteristic data and in combination with the characteristic weight data corresponding to the disease to be evaluated, determine a single characteristic sequence set; wherein, the single characteristic sequence includes the weighted sum result of each characteristic and the corresponding characteristic weight in each day within the monitoring period;

[0009] Input the single characteristic sequence set into a trained risk assessment model to obtain the outbreak risk assessment result of the disease to be evaluated within the preset monitoring period.

[0010] In an exemplary embodiment of the present disclosure, the method further includes: 1]

[0011] According to the basic reproduction number of the infectious disease corresponding to the disease to be evaluated, perform eigenvalue weighting processing on the single characteristic sequence set to obtain an enhanced single characteristic sequence set, and use this enhanced single characteristic sequence set as the input of the risk assessment model.

[0012] In an exemplary embodiment of the present disclosure, the obtaining the characteristic data of a preset monitoring period includes:

[0013] Collect the monitoring data of the target area within the preset monitoring period duration;

[0014] Preprocess the monitoring data according to the feature set corresponding to the disease to be evaluated to obtain the feature data.

[0015] In an exemplary embodiment of the present disclosure, the method further includes: pre-determining the feature weights corresponding to the disease to be evaluated according to the historical data to be evaluated, including:

[0016] Collect the historical data corresponding to the disease to be evaluated; wherein, the historical data includes the feature statistical data corresponding to each feature in several consecutive transmission cycles before the outbreak date of the disease to be evaluated;

[0017] Determine the mean and standard deviation corresponding to each feature in each transmission cycle;

[0018] Input the mean and standard deviation corresponding to the feature into the trained feature contribution degree model to respectively obtain the weights corresponding to the mean and standard deviation of the feature;

[0019] Configure the sum of the weights corresponding to the mean and standard deviation of each feature as the feature weight.

[0020] In an exemplary embodiment of the present disclosure, after determining the weights corresponding to each feature, the method further includes:

[0021] In response to the weight adjustment instruction, adjust the weight value corresponding to the selected feature based on the relevance between the feature and the disease to be evaluated.

[0022] In an exemplary embodiment of the present disclosure, after collecting the historical data corresponding to the disease to be evaluated, the method further includes:

[0023] In response to the feature adjustment instruction, select the features of the target type in the historical data based on the relevance between the disease to be evaluated and the environmental features and features of the target area for calculating the weights of these features.

[0024] In an exemplary embodiment of the present disclosure, the risk assessment model is a risk assessment model based on the logistic regression algorithm or a risk assessment model based on the long short-term memory network algorithm; the historical data includes the early warning intensity parameters corresponding to each transmission cycle;

[0025] The method further includes:

[0026] Determine a single feature sequence set based on the feature data corresponding to the features in the historical data in combination with the feature weights; wherein, the single feature sequence includes the weighted sum result of each feature and the corresponding feature weight in each day of the transmission cycle;

[0027] Construct training samples based on the single feature sequence set and the corresponding early warning intensity parameters, and use the training samples to train the risk assessment model.

[0028] According to a second aspect of the present disclosure, there is provided a disease monitoring and early warning device, including:

[0029] A data acquisition module for acquiring feature data of a preset monitoring period;

[0030] A single feature sequence set calculation module for determining a single feature sequence set by combining the feature weights corresponding to the disease to be evaluated and the feature data;

[0031] An evaluation result output module for inputting the single feature sequence set into the trained evaluation model to obtain the transmission risk assessment result of the disease to be evaluated.

[0032] According to a third aspect of the present disclosure, there is provided a storage medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned disease monitoring and early warning method is implemented.

[0033] According to a fourth aspect of the present disclosure, there is provided an electronic device, including:

[0034] A processor; and

[0035] A memory for storing executable instructions of the processor;

[0036] Wherein, the processor is configured to execute the above-mentioned disease monitoring and early warning method by executing the executable instructions.

[0037] In the disease monitoring and early warning method provided by an embodiment of the present disclosure, by collecting feature data within the duration of a preset monitoring period, a single feature sequence set is constructed using the feature data and the feature weight data corresponding to each feature calculated in advance, and then the risk assessment model is used to determine the outbreak risk assessment result of the disease to be evaluated in this monitoring period based on the single feature sequence set. It is possible to use multiple observable features as indicators to give early warnings for infectious diseases and improve the accuracy of early warnings for the outbreak risks of infectious diseases.

[0038] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure. Obviously, the accompanying drawings in the following description are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0040] Figure 1 A schematic diagram schematically showing a disease monitoring and early warning method in an exemplary embodiment of the present disclosure;

[0041] Figure 2 A schematic diagram schematically showing a system architecture in an exemplary embodiment of the present disclosure;

[0042] Figure 3 A schematic diagram schematically showing a risk assessment model training method in an exemplary embodiment of the present disclosure;

[0043] Figure 4 A schematic diagram schematically showing the composition of a disease monitoring and early warning device in an exemplary embodiment of the present disclosure;

[0044] Figure 5 A schematic diagram schematically showing the composition of a terminal device in an exemplary embodiment of the present disclosure;

[0045] Figure 6 A schematic diagram schematically showing the composition of a storage medium in an exemplary embodiment of the present disclosure. Detailed implementation manners

[0046] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art. The features, structures, or characteristics described can be combined in any suitable manner in one or more embodiments.

[0047] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can 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.

[0048] The block diagrams shown in the drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0049] The flowcharts shown in the drawings are only illustrative and do not necessarily include all content and operations / steps, nor are they necessarily executed in the described order. For example, some operations / steps can be decomposed, while some operations / steps can be combined or partially combined. Therefore, the actual execution order may change according to the actual situation.

[0050] In the related art, according to the change in the number of inpatients in a certain area during a certain period, combined with various early warning algorithms, such as the moving percentile method, the cumulative sum control chart method, etc., it is possible to judge whether there is a possibility of an infectious disease outbreak, so as to achieve the purpose of early warning. Taking influenza as an example, the moving percentile method can be used to model the number of inpatients at different times, and different percentile P values are used as candidate early warning threshold values to construct a conventional influenza model to achieve the purpose of early warning. However, using the number of inpatients as an early warning indicator, the infected people may choose to stay at home / take medicine / seek medical treatment. In order to prevent false alarms, different early warning models often only alarm after the number of inpatients reaches a certain trend, either large or small; this also indicates that a certain scale of infection has been reached, with a certain lag. Moreover, due to the relatively single indicator relied on, the early warning of the epidemic situation cannot achieve the purpose of being timely and effective, and there may be a risk of false alarms.

[0051] In order to solve the technical problems existing in the prior art, in this exemplary embodiment, a disease monitoring and early warning method is first provided, which can be applied to analyze, evaluate, and accurately warn of the outbreak of infectious disease epidemics. Referring to Figure 1 as shown, the above-mentioned disease monitoring and early warning method may include the following steps:

[0052] Step S11, obtaining characteristic data for a preset monitoring period;

[0053] Step S12, determining a single characteristic sequence set according to the characteristic data in combination with the characteristic weight data corresponding to the disease to be evaluated; wherein, the single characteristic sequence includes the weighted summation result of each characteristic and the corresponding characteristic weight in each day within the monitoring period;

[0054] Step S13, inputting the single characteristic sequence set into a trained risk assessment model to obtain the outbreak risk assessment result of the disease to be evaluated within the preset monitoring period.

[0055] The disease monitoring and early warning method provided by this exemplary embodiment collects feature data within a preset monitoring cycle duration, constructs a single-feature sequence set using these feature data and the feature weight data corresponding to each feature calculated in advance, and then uses a risk assessment model to determine the outbreak risk assessment result of the disease to be evaluated during this monitoring cycle based on this single-feature sequence set. It can use multiple observable features as indicators to give early warnings for infectious diseases and improve the accuracy of early warnings for the outbreak risks of infectious diseases.

[0056] Next, each step of the disease monitoring and early warning method in this exemplary embodiment will be described in more detail in conjunction with the accompanying drawings and embodiments.

[0057] In this exemplary embodiment, refer to Figure 2 As shown, a system architecture is provided, which can be used to execute and implement the above-mentioned disease monitoring and early warning method. Refer to Figure 2 As shown, the system architecture may include a terminal device 201, a network 202, and a server 203. Among them, the terminal device 201 may be one or more of a smart phone, a tablet computer, a desktop computer, and a portable computer; the network 202 may be a medium for providing a communication link between the terminal device and the server. The network 202 may include various connection types, such as a wired communication link, a wireless communication link, and so on. It should be understood that Figure 2 The numbers of terminal devices, networks, and servers in

[0058] In this exemplary embodiment, the above-mentioned disease monitoring and early warning method may pre-train a risk assessment model. Specifically, refer to Figure 3 As shown, the training method of the risk assessment model may include:

[0059] Step S31, collect historical data corresponding to the disease to be evaluated; wherein, the historical data includes the feature statistical data corresponding to each feature in several consecutive transmission cycles before the outbreak date of the disease to be evaluated;

[0060] Step S32, determine the mean and standard deviation corresponding to each feature in each transmission cycle;

[0061] Step S33: Input the mean and standard deviation corresponding to the feature into the trained feature contribution model to obtain the weights corresponding to the mean and standard deviation of the feature respectively; and configure the sum of the weights corresponding to the mean and standard deviation of each feature as the feature weight.

[0062] Step S34: Determine the single-feature sequence set based on the feature data corresponding to the historical data in combination with the feature weight; wherein, the single-feature sequence includes the weighted summation result of each feature and the corresponding feature weight in each day within the propagation period.

[0063] Step S35: Construct a training sample according to the single-feature sequence set and the corresponding early warning intensity parameter, and use this training sample to train the risk assessment model.

[0064] In this exemplary embodiment, in the above step S31, the historical data of a certain disease to be evaluated can be the historical data of the first n days before the outbreak of the disease in a region. Wherein, n is a positive integer. For example, if the outbreak date of influenza A in City A is December 1st, then the continuous n days of historical data before December 1st can be extracted. Or, the value of n can also be determined according to the propagation period of the disease. For example, when the propagation period is k days, the historical data of an integer multiple of k days can be selected; for example, the historical data of 2k, 3k, 4k, 5k or 6k days can be collected.

[0065] Specifically, for the current area to be early warned, it is necessary to select the types of infectious diseases that need to be early warned, so as to obtain the historical outbreak time point records of this area for this type of infectious disease. Assume that k is the propagation period of a certain type of infectious disease obtained based on past research statistics. If the period is not clear, it can be set according to the propagation period of conventional infectious diseases, for example, set to 7.

[0066] Among them, the historical data can include historical environmental data and historical medical data; through induction and statistics of the historical data, the feature statistical data corresponding to the disease to be evaluated is obtained. For different infectious diseases, their corresponding features can also be different, so that risk assessment models for different disease types can be trained respectively. Specifically, corresponding to the historical environmental data and historical medical data included in the historical data, the features corresponding to a disease type can also include environmental features and medical features. For example, environmental features can include: the average temperature of the day, whether it is rainy or snowy, the average humidity of the day, etc. Medical features can include: the sales volume of drug A, the number of consultations in a certain department, the number of a certain type of diagnosis results, and so on. In addition, for a disease type, different environmental features can also be selected for different regions.

[0067] For example, after obtaining historical data, statistics and summarization can be performed on the data of each feature. For example, as shown in Table 1, the feature statistical data of each feature can be counted for each day.

[0068]

[0069] Table 1

[0070] In some exemplary embodiments, when collecting historical data and determining the features corresponding to a disease in advance, the historical data corresponding to these features can be directly extracted; or, after obtaining all the historical data, the features and the corresponding feature statistical data can be screened according to the disease assessment requirements.

[0071] For the epidemics that broke out in different historical time periods, for the data sequence of the above-mentioned historical data collected, intercept the sequence corresponding to the k days before the day of the starting point of the outbreak time to the day before, and set the outcome variable corresponding to this sequence, that is, the warning intensity, to 2, indicating that the epidemic will break out soon; for the time series data collected from 2k days before the outbreak time point to k days, set the warning intensity corresponding to this sequence to 1; for the time series data collected from 3k days before the outbreak time point to 2k days, set the warning intensity corresponding to this sequence to 1 / 2; for the time series data collected from 4k days before the outbreak time point to 3k days, set the warning intensity corresponding to this sequence to 1 / 4; for the time series data collected from 5k days before the outbreak time point to 4k days, set the warning intensity corresponding to this sequence to 1 / 8. The setting of the warning intensity is a relative value, and the purpose is to distinguish the warning levels of different sequences, so as to provide data for the subsequent calculation of the weights of different warning features. The numbers here are only for illustration, but should conform to the above multiple relationship. In addition, for other time periods without an outbreak of the epidemic, intercept in sequence according to k days, and the data less than k days can be discarded; and set the outcome variables corresponding to these sequences to 0, indicating that there is no outbreak of the epidemic. For the short-term continuously breaking out epidemic, the assignment of the outcome variable Y may be repeated. For example, the current time series is in the time period from 4k days before the outbreak time point to 3k, and is also in the time period from k days before another epidemic to the day before; then according to the above logic, select the outcome variable with the larger assignment as the outcome variable of the current sequence.

[0072] Based on the above content, time series data with a time length of k for different features will be obtained. Each sequence corresponds to the calculated outcome variable value Y. The sequence data shown in Table 2 includes the statistical results of each feature for each day within a transmission cycle and the corresponding warning level; as the complete feature data for a transmission cycle or a monitoring cycle.

[0073]

[0074] Table 2

[0075] In this exemplary embodiment, in step S32 described above, for the feature Fi of each dimension, the time series data within k days of it can be converted into the mean Mean_Fi and the standard deviation Std_Fi. The original time series data will be flattened into non-time series data, that is, the mean and standard deviation of each feature within k days, recorded as X, and the corresponding warning intensity Y. The calculation results can be as shown in Table 3, including all the flattened time series data and the corresponding outcome variables

[0076]

[0077] Table 3

[0078] In this exemplary embodiment, in step S33 described above, the above feature contribution model can adopt a gradient boosting decision tree model to construct a regression problem of training data with the mean and standard deviation of each feature within k days as the input X and the final warning intensity Y as the output, and train to obtain a tree model.

[0079] Specifically, the training of the feature contribution model based on the gradient boosting decision tree model can include: initializing the first tree for fitting, and the formula can include:

[0080]

[0081] where x is the mean and standard deviation, y is the warning intensity; γ represents the set of function parameters to be fitted corresponding to the current tree model F0(x).

[0082] A total of M trees are constructed for residual fitting, traversing from m = 1 to m = M.

[0083] Then, calculate the gradient vector of the loss function on the existing model: That is, the gradient of the existing model, where L represents the loss function.

[0084] According to this gradient value, obtain the residual R that the current model needs to fit im , and the calculation formula can include:

[0085]

[0086] where i represents the i-th sample.

[0087] According to the calculation result of this residual, and the current sample point set X, form a training set {(X i , R im ), i = 1,..., n.

[0088] Based on this training set, a weak classifier is trained. This classifier is the basic GBRT (Gradient Boosting Regression Tree), and the training method is the same as the method for fitting the first tree. The obtained model is denoted as h m (x i ). Its calculation formula can include:

[0089]

[0090] To obtain the weight of this new base classifier in the entire model after the m-th iteration, and to find the optimal weight, the new base classifier is added to the original entire model combination, denoted as F m-1 (x i ) + γ × h m (x i ). To obtain the weight parameter before h, use the line search method, put it into the loss function, and find the optimal γ m . Among them,

[0091]

[0092] The updated overall model can include: F m (x) = F m,1 (x i ) + γ × h m (x i )

[0093] By analyzing the overall model F m (x), calculate the degree of variance reduction when splitting the non-leaf nodes constructed by all trees, and finally calculate the average value of the importance of each feature for all nodes to obtain the weight of each feature. Specifically, it can include: for each feature M, specifically, that is, Mean_Fi and Std_Fi corresponding to all features i; for each tree in F m (x), find the branch nodes of the non-leaf nodes formed by M, and calculate the importance statistic of the split feature M corresponding to this node. Specifically, that is, calculate the variance of all data of the current node before splitting, subtract the sum of the variances of the left and right binary trees after splitting. Each variance value is multiplied by the corresponding weight (the number of samples of the branch). For example, if the variance before splitting is S1, the variances of the left and right binary trees after splitting are S2 and S3, the sample size of the left binary tree after splitting is A, and the sample size of the right binary tree after splitting is B, then the importance statistic of the split feature M corresponding to this node = S1 - (S2 × A / (A + B) + S3 × B / (A + B)). Calculate the sum of the importance statistics of the split feature M corresponding to all nodes of the current tree, and record it as S. For F mAfter traversing each tree in (x), the importance statistics of the splitting feature M corresponding to each tree are obtained. After performing the above process on all features, the importance statistics obtained for each feature are min-max normalized, that is, normalized, and finally the normalized weight values corresponding to each input feature are obtained, as shown in Table 4.

[0094] Mean_F1 Std_F1 … … Mean_Fn Std_Fn Weight value Weight value … … Weight value Weight value

[0095] Table 4

[0096] Take the sum of the mean Mean_Fi and the standard deviation Std_Fi as the weight score Si of the current feature Fi. There is no need to reflect the respective weights of Mean_Fi and Std_Fi. For a feature, directly use the mean weight of the feature + the standard deviation weight of the feature to obtain the overall weight of the feature Fi.

[0097] In this exemplary embodiment, in the above step S34, the obtained weights Si can be directly used to perform weighted summation on all features to obtain a single-feature time series data that combines all feature information. The single-feature value obtained after combination can be expressed as:

[0098]

[0099] For the sequence data set of multiple features with a time length of k that has not been flattened as described above, the above weighted summation is performed on all of them. Finally, a set of single-feature time series with a time length of k will be obtained; different features are fused and a single feature is generated. Each sequence corresponds to an early warning intensity Y, and the specific representation of the data of each sequence is shown in Table 5.

[0100] <![CDATA[T1]]> <![CDATA[T2]]> … <![CDATA[T k > Warning intensity <![CDATA[Combined feature F sum > 2.3 2.6 2.4 2.7 1 / 2

[0101] Table 5

[0102] Based on the above, after obtaining the single-feature sequence set, the data inside the sequence can also be enhanced. Specifically, the characteristics of the epidemic spread presenting an exponential function can also be considered. For example, the weighted moving average method is used to enhance the single-feature time series.

[0103] For example, it is known that the basic reproduction number R of an infectious disease can be obtained more clearly according to historical research. Although the spread of the infectious disease has shown an exponential function form, in the early stage of the epidemic, the spread trend is not obvious. Therefore, based on the current transmission cycle k and the basic reproduction number R of the infectious disease, we can assign weights to the k corresponding feature F sum values at time T1 to T of the sequence. The methods to enhance the increasing trend of the exponential function in the sequence can include:

[0104]

[0105] The finally obtained weighted sequence X new , taking the data in the above example as an example, as shown in Table 6, it can be expressed as:

[0106]

[0107] Alternatively, in some exemplary embodiments, for the data augmentation processing of the single-feature sequence set, an arithmetic progression form can also be used, calculated based on the propagation period k and the basic reproduction number R of the infectious disease.

[0108] In this exemplary embodiment, in the above step S35, for sequences of length k in different periods, after performing the above-mentioned weighting process on each period of the sequence, the data-augmented sequence X can be obtained. new , and then according to the outcome variable Y corresponding to the current sequence obtained in the above steps, that is, the early warning intensity value, a new data-augmented training set can be constructed. Using the long short-term memory network LSTM, a regression task is constructed and the constructed new training set [X new , Y] is used to train the model.

[0109] Specifically, the long short-term memory network model is controlled by an input gate, an output gate, and a forget gate for the cell memory state. The gate values generally range from 0 to 1. Therefore, the three gates use the Sigmoid function as the activation function, and the output state uses the Tanh activation function.

[0110] The weighted single-feature time series X of length k new is input into the LSTM model, and the outcome variable Y corresponding to the current sequence, that is, the early warning intensity value, is used as the output of the LSTM to train the model to obtain the model LSTM(X). This model is the evaluation model for subsequent early warning prediction of the current infectious disease.

[0111] In some exemplary embodiments, other time series models can also be used to train the risk assessment model; for example, the CNN model, ARIMA (Autoregressive Integrated Moving Average model); or a regression tree model can be adopted.

[0112] In step S11, the feature data of the preset monitoring period is obtained.

[0113] In this exemplary embodiment, the above step S11 may include:

[0114] Step S111, collecting the monitoring data of the target area within the duration of the preset monitoring period;

[0115] Step S112, preprocess the monitoring data according to the feature set corresponding to the disease to be evaluated to obtain the feature data.

[0116] Specifically, the target area may be an area to be warned where there is a risk of infectious disease outbreak; after determining the disease type of the disease to be evaluated, the corresponding transmission cycle can be determined; the above monitoring cycle may include the duration of at least one transmission cycle; for example, the data of the k days before the current day can be used as the monitoring data. As in the above embodiments, the monitoring data may include environmental data and medical data. After collecting the monitoring data, the monitoring data can be statistically analyzed and sorted to obtain the feature data corresponding to the disease to be evaluated. Among them, the feature data may include environmental features and medical features, as well as the feature statistical data corresponding to each dimension feature. As shown in Table 1, the feature statistical data corresponding to each dimension feature in each day of the monitoring cycle is obtained after statistics.

[0117] In some exemplary embodiments, the above method may further include: in response to a feature adjustment instruction, select features of a target type from the historical data based on the relevance between the disease to be evaluated and the environmental features and features of the target area for calculating the weights of these features.

[0118] For example, for the disease to be evaluated, when organizing the monitoring data, new features can be added or existing features can be deleted manually; thereby realizing the update of the features and using these updated features for calculation.

[0119] In step S12, determine a single feature sequence set according to the feature data in combination with the feature weight data corresponding to the disease to be evaluated; wherein, the single feature sequence includes the weighted sum result of each feature and the corresponding feature weight in each day of the monitoring cycle.

[0120] In the present exemplary embodiment, according to the feature weights corresponding to each feature of the disease to be evaluated that have been determined, sum the calculation sets of each feature combined with the corresponding weights in each day of the monitoring cycle to determine the single feature sequence set. The calculation formula may include:

[0121]

[0122] Where Si is the weight and Fi is the feature value of the feature. The calculation results are shown in Table 5.

[0123] In some exemplary embodiments, the above method may further include: in response to a weight adjustment instruction, adjust the weight value corresponding to the selected feature based on the relevance between the feature and the disease to be evaluated.

[0124] For example, for the existing feature weights of each feature, before calculating the single-feature sequence set, the weight values corresponding to each feature can also be adjusted according to the actual situation; for example, the weight values of some features can be increased or the weight values of some features can be decreased according to the actual changes in infectious diseases and the environment.

[0125] In the embodiment of this example, preferably, after obtaining the single-feature sequence set, the single-feature sequence set can also be weighted with eigenvalue according to the basic reproduction number of the infectious disease corresponding to the disease to be evaluated, so as to obtain a single-feature sequence set after data enhancement, and use this single-feature sequence set after data enhancement as the input of the risk assessment model.

[0126] Specifically, according to the basic reproduction number R of the infectious disease corresponding to the disease to be evaluated, based on the transmission period k and the basic reproduction number R of the disease to be evaluated, the feature F corresponding to T1 to Tk within the monitoring period sum values are weighted, and its calculation formula can include:

[0127]

[0128] In step S13, the single-feature sequence set is input into the trained risk assessment model to obtain the outbreak risk assessment result of the disease to be evaluated within the preset monitoring period.

[0129] In the embodiment of this example, the finally enhanced input sequence X in is obtained, and then input into the trained LSTM model to obtain, for the current moment, in the current area to be warned, the warning value Y of the current disease to be evaluated, Y = LSTM(X in ); and use this warning value as the outbreak risk assessment result of the disease to be evaluated.

[0130] For example, when Y>2, it can indicate that the epidemic is about to break out soon and is very dangerous; when 1<Y<2, it can indicate that there is a risk of epidemic outbreak within k days; when 0.5<Y<1, it can indicate that there is a risk of epidemic outbreak within 2k to k days; when 0.25<Y<0.5, it can indicate that there is a risk of epidemic outbreak within 3k to 2k days; when 0.125<Y<0.25, it can indicate that there is a risk of epidemic outbreak within 4k to 3k days; when Y<0.125, it can indicate that there is no risk of epidemic for the time being.

[0131] The disease monitoring and early warning method provided in the exemplary embodiments of the present disclosure configures multiple countable features for different diseases as evaluation indicators, and collects data for each day within the monitoring period to achieve fine-grained assignment. Thus, in addition to the binary result of whether to give an early warning, it can more finely predict the time period of the epidemic outbreak. By means of automatic weighting, a comprehensive index is summarized. While ensuring the interpretability of the weighting process, it allows manual adjustment of the index weights in a targeted manner, and combines data augmentation means to complete more accurate early warning for different disease types within a fixed area.

[0132] Further, referring to Figure 4 As shown, in the embodiment of this example, a disease monitoring and early warning device 40 is further provided, including: a data acquisition module 401, a single feature sequence set calculation module 402, and an evaluation result output module 403. Among them,

[0133] The data acquisition module 401 can be used to obtain the feature data of the preset monitoring period;

[0134] The single feature sequence set calculation module 402 is used to determine a single feature sequence set in combination with the feature weights corresponding to the disease to be evaluated and the feature data.

[0135] The evaluation result output module 403 can be used to input the single feature sequence set into the trained evaluation model to obtain the evaluation result of the transmission risk of the disease to be evaluated.

[0136] In some exemplary embodiments, the above device 40 may further include: a data augmentation processing module.

[0137] The data augmentation processing module can be used to perform eigenvalue weighting processing on the single feature sequence set according to the basic reproduction number of the infectious disease corresponding to the disease to be evaluated, so as to obtain a data-augmented single feature sequence set, and use this data-augmented single feature sequence set as the input of the risk assessment model.

[0138] In some exemplary embodiments, the data acquisition module 401 can be used to collect the monitoring data of the target area within the preset monitoring period duration; preprocess the monitoring data according to the feature set corresponding to the disease to be evaluated to obtain the feature data.

[0139] In some exemplary embodiments, the above device 40 may further include: a weight preprocessing module.

[0140] The weight preprocessing module can be used to estimate the feature statistical data corresponding to each feature in a number of consecutive transmission cycles before the disease outbreak date; determine the mean and standard deviation corresponding to each feature in each transmission cycle; input the mean and standard deviation corresponding to the feature into the trained feature contribution model to respectively obtain the weights corresponding to the mean and standard deviation of the feature; and configure the sum of the weights corresponding to the mean and standard deviation of each feature as the feature weight.

[0141] In some exemplary embodiments, the above-mentioned device 40 may further include: a weight value adjustment module.

[0142] The weight value adjustment module can be used to respond to a weight adjustment instruction and adjust the weight value corresponding to the selected feature based on the relevance between the feature and the disease to be evaluated.

[0143] In some exemplary embodiments, the above-mentioned device 40 may further include: a weight type adjustment module.

[0144] The weight type adjustment module can be used to respond to a feature adjustment instruction and select features of a target type from the historical data based on the relevance between the disease to be evaluated and the environmental features and features of the target area for calculating the weights of these features.

[0145] In some exemplary embodiments, the risk assessment model is a risk assessment model based on a logistic regression algorithm or a risk assessment model based on a long short-term memory network algorithm; the historical data includes the early warning intensity parameters corresponding to each transmission cycle;

[0146] The above-mentioned device 40 may further include: a risk assessment model pre-training module.

[0147] The risk assessment model pre-training module can be used to determine a single feature sequence set based on the feature data corresponding to the features in the historical data in combination with the feature weights; wherein, the single feature sequence includes the weighted sum result of each feature and the corresponding feature weight in each day of the transmission cycle; construct a training sample according to the single feature sequence set and the corresponding early warning intensity parameter, and use the training sample to train the risk assessment model.

[0148] The specific details of each module in the above-mentioned disease monitoring and early warning device have been described in detail in the corresponding disease monitoring and early warning method, so they will not be repeated here.

[0149] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, such a division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more of the above-described modules or units can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0150] In an exemplary embodiment of the present disclosure, a computer system capable of implementing the above method is also provided.

[0151] Those skilled in the art can understand that various aspects of the present invention can be implemented as a system, a method, or a program product. Therefore, various aspects of the present invention can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "system" here.

[0152] Next, refer to Figure 5 to describe the terminal device 500 according to this embodiment of the present invention. Figure 5 The composition of the displayed terminal device 500 is only an example and should not impose any limitations on the functions and usage scope of the embodiments of the present invention.

[0153] As Figure 5 shown, the components of the terminal device 500 may include, but are not limited to: at least one of the above-mentioned processing units 510, at least one of the above-mentioned storage units 520, and a bus 530 connecting different system components (including the storage unit 520 and the processing unit 510).

[0154] Among them, the storage unit stores program code, and the program code can be executed by the processing unit 510, so that the processing unit 510 executes 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 510 can execute the steps as Figure 1 shown in.

[0155] The storage unit 520 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 5201 and / or a cache storage unit 5202, and may further include a read-only storage unit (ROM) 5203.

[0156] The storage unit 520 may also include a program / utility 5204 having a set (at least one) of program modules 5205. Such program modules 5205 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0157] The bus 530 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus structures.

[0158] The terminal can also communicate with one or more external devices 300 (such as a keyboard, a pointing device, a Bluetooth device, etc.), and can also communicate with one or more devices that enable a user to interact with the terminal, and / or communicate with any device that enables the terminal to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication can be carried out through the input / output (I / O) interface 550. Also, the terminal can communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the network adapter 560. As shown in the figure, the network adapter 560 communicates with other modules of the terminal through the bus 530. The processing unit 510 is connected to the display unit 540 through the bus 530. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the terminal, 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.

[0159] Through the description of the above embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented by software, or can be implemented by a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure 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 several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0160] In an exemplary embodiment of the present disclosure, there is also provided a computer-readable storage medium, on which a program product capable of implementing the above-described method of this specification is stored. In some possible implementation manners, various aspects of the present invention can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of this specification.

[0161] Reference Figure 6 As shown, a program product 600 for implementing the above method according to an embodiment of the present invention is described. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0162] The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. A readable storage medium can, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection having 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0163] A computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries the readable program code. Such a propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable signal medium can also be any readable medium other than the readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0164] The program code contained on the readable medium can be transmitted by any appropriate medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination of the above.

[0165] 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 also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or it may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0166] In addition, the above-mentioned drawings are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present invention, and are not for limiting purposes. It is easy to understand that the processes shown in the above-mentioned drawings do not indicate or limit the chronological order of these processes. Additionally, it is also easy to understand that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0167] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the claims.

[0168] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A disease monitoring and early warning method, characterized in that, Including: Obtain the characteristic data of the preset monitoring period; According to the characteristic data and the characteristic weight data corresponding to the disease to be evaluated, determine a single characteristic sequence set; wherein, the single characteristic sequence includes the weighted sum result of each characteristic and the corresponding characteristic weight in each day within the monitoring period; Input the single characteristic sequence set into the trained risk assessment model to obtain the outbreak risk assessment result of the disease to be evaluated within the preset monitoring period; the risk assessment model is a risk assessment model based on the logistic regression algorithm or a risk assessment model based on the long short-term memory network algorithm; Wherein, the training method of the risk assessment model includes: Collect the historical data corresponding to the disease to be evaluated; wherein, the historical data includes the characteristic statistical data corresponding to each characteristic in several consecutive transmission periods before the outbreak day of the disease to be evaluated and the early warning intensity parameters corresponding to each transmission period; Determine the mean and standard deviation corresponding to each characteristic in each transmission period; Input the mean and standard deviation corresponding to the characteristic into the trained characteristic contribution degree model to respectively obtain the weights corresponding to the mean and standard deviation of the characteristic; configure the sum of the weights corresponding to the mean and standard deviation of each characteristic as the characteristic weight; Based on the characteristic data corresponding to the characteristics in the historical data and the characteristic weight, determine a single characteristic sequence set; wherein, the single characteristic sequence includes the weighted sum result of each characteristic and the corresponding characteristic weight in each day within the transmission period; Construct a training sample according to the single characteristic sequence set and the corresponding early warning intensity parameter, and use the training sample to train the risk assessment model; Wherein, the characteristic contribution degree model adopts a gradient boosting decision tree model.

2. The disease monitoring and early warning method according to claim 1, wherein The method further includes: Perform eigenvalue weighting processing on the single characteristic sequence set according to the basic reproduction number of the infectious disease corresponding to the disease to be evaluated to obtain an enhanced single characteristic sequence set after data enhancement, and use the enhanced single characteristic sequence set after data enhancement as the input of the risk assessment model.

3. The disease monitoring and early warning method according to claim 1, characterized in that The obtaining of the characteristic data of the preset monitoring period includes: Collect the monitoring data of the target area within the preset monitoring period duration; Preprocess the monitoring data according to the characteristic set corresponding to the disease to be evaluated to obtain the characteristic data.

4. The disease monitoring and early warning method according to claim 1, wherein After determining the weight corresponding to each characteristic, the method further includes: In response to the weight adjustment instruction, adjust the weight value of the selected characteristic based on the relevance between the characteristic and the disease to be evaluated.

5. The disease monitoring and early warning method according to claim 1, characterized in that After collecting the historical data corresponding to the disease to be evaluated, the method further includes: In response to the characteristic adjustment instruction, select the characteristics of the target type in the historical data based on the relevance between the disease to be evaluated and the environmental characteristics and characteristics of the target area for calculating the weights of these characteristics.

6. A disease monitoring and early warning device, characterized in that, Including: A data collection module for obtaining the characteristic data of the preset monitoring period; A single characteristic sequence set calculation module for determining a single characteristic sequence set by combining the characteristic weight corresponding to the disease to be evaluated and the characteristic data; An evaluation result output module is configured to input the single-feature sequence set into a trained evaluation model to obtain an evaluation result of the transmission risk of the disease to be evaluated; the risk assessment model is a risk assessment model based on a logistic regression algorithm or a risk assessment model based on a long short-term memory network algorithm; Wherein, the training method of the risk assessment model includes: collecting historical data corresponding to the disease to be evaluated; wherein, the historical data includes feature statistical data corresponding to each feature and early warning intensity parameters corresponding to each transmission cycle within a plurality of consecutive transmission cycles before the outbreak date of the disease to be evaluated; determining the mean and standard deviation corresponding to each feature within each transmission cycle; inputting the mean and standard deviation corresponding to the feature into a trained feature contribution model to respectively obtain the weights corresponding to the mean and standard deviation of the feature; configuring the sum of the weights corresponding to the mean and standard deviation of each feature as a feature weight; determining a single-feature sequence set based on the feature data corresponding to the historical data in combination with the feature weight; wherein, the single-feature sequence includes the weighted summation result of each feature and the corresponding feature weight on each day within the transmission cycle; constructing a training sample according to the single-feature sequence set and the corresponding early warning intensity parameter, and training the risk assessment model by using the training sample; wherein, the feature contribution model adopts a gradient boosting decision tree model.

7. A storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the disease monitoring and early warning method according to any one of claims 1 to 5.

8. A terminal device, characterized in that, Comprising: A processor; And A memory for storing executable instructions of the processor; Wherein, the processor is configured to execute the disease monitoring and early warning method according to any one of claims 1 to 5 by executing the executable instructions.

Images