Epidemic disease prediction method, device, equipment and medium

Abstract

The invention discloses an epidemic disease prediction method and device, equipment and a medium. The method comprises the following steps: acquiring drug sales data, drug attributes and drug use purposes of at least one drug in a historical time period; predicting drug sales data of each drug in a future time period according to the drug sales data of each drug in the historical time period; according to the medicine sales data, the medicine attributes and the medicine use purposes of all the medicines, the epidemic disease pointing relation of all the medicines is determined; and according to the epidemic disease pointing relationship of each medicine and the medicine sales data of each medicine in the future time period, determining the spreading space-time information of the epidemic disease pointed by each medicine. According to the embodiment of the invention, the accuracy of epidemic prediction can be improved.

Description

Technical Field

[0001] The present invention relates to the field of epidemic control technology, and in particular to an epidemic prediction method, device, equipment and medium. Background Art

[0002] With the acceleration of globalization and the continuous development of human society, the speed and scope of epidemic spread are increasing, posing huge challenges to public health and social economy. Therefore, it is necessary to timely and effectively predict the occurrence, development and spread of epidemics to reduce the risks and social impacts of epidemic spread.

[0003] Currently, epidemic predictions rely on traditional statistical models and manual analysis to predict future disease transmission trends.

[0004] However, the above prediction methods have the disadvantages of low efficiency and low prediction accuracy. Summary of the invention

[0005] The present invention provides an epidemic prediction method, device, equipment and medium. The embodiments of the present invention can improve the accuracy of epidemic prediction.

[0006] In a first aspect, an embodiment of the present invention provides an epidemic prediction method, the method comprising:

[0007] Obtain drug sales data, drug attributes, and drug usage for at least one drug in a historical time period;

[0008] Based on the drug sales data of each drug in the historical time period, predict the drug sales data of each drug in the future time period;

[0009] Determine the epidemiological relationship of each drug based on its sales data, drug properties and drug usage;

[0010] According to the epidemiological directional relationship of each drug and the drug sales data of each drug in the future time period, the spatiotemporal information of the spread of the epidemic pointed by each drug is determined.

[0011] In a second aspect, an embodiment of the present invention further provides an epidemic prediction device, the device comprising:

[0012] A data acquisition module, used to acquire drug sales data, drug attributes and drug usage of at least one drug in a historical time period;

[0013] A prediction module is used to predict the drug sales data of each drug in a future time period based on the drug sales data of each drug in a historical time period;

[0014] A relationship determination module, for determining the epidemiological orientation relationship of each drug based on drug sales data, drug attributes and drug usage of each drug;

[0015] The information determination module is used to determine the spatiotemporal information of the epidemic spread pointed by each drug based on the epidemic pointing relationship of each drug and the drug sales data of each drug in the future time period.

[0016] In a third aspect, an embodiment of the present invention further provides an epidemic prediction device, the epidemic prediction device comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the epidemic prediction method of any embodiment of the present invention.

[0020] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to implement the epidemic prediction method of any embodiment of the present invention when executed by a processor.

[0021] The technical solution of the embodiment of the present invention can achieve a comprehensive understanding of drugs by acquiring drug sales data, drug properties and drug usage in historical time periods, thereby providing a data basis for subsequent predictions; by analyzing historical drug sales data, it is possible to predict future drug sales trends and provide a reference for epidemic predictions; by combining drug sales data, properties and uses, it is possible to analyze the potential relationship between drugs and epidemics; by combining the directional relationship between drugs and epidemics and future sales data, it is possible to predict the spatiotemporal characteristics of epidemic transmission, thereby improving the accuracy of epidemic predictions.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 A flowchart of an epidemic prediction method provided by an embodiment of the present invention;

[0025] Figure 2 A flowchart of an epidemic prediction method provided by an embodiment of the present invention;

[0026] Figure 3 A schematic diagram of the network structure of a multi-head self-attention expert network provided by an embodiment of the present invention;

[0027] Figure 4 A schematic diagram of a network structure of a multi-scale feature pyramid provided by an embodiment of the present invention;

[0028] Figure 5 A schematic diagram of the structure of an epidemic prediction device provided by an embodiment of the present invention;

[0029] Figure 6 A schematic diagram of the structure of an epidemic prediction device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] In the technical solution of the embodiment of the present invention, the acquisition, storage and application of the drug sales data, user purchase records and user purchase behavior involved all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0033] Figure 1A flowchart of an epidemic prediction method provided by an embodiment of the present invention. The embodiment of the present invention is applicable to the situation of epidemic prediction, and the method can be executed by an epidemic prediction device, which can be implemented in the form of hardware and / or software.

[0034] See also Figure 1 The epidemic prediction methods shown include:

[0035] S101. Obtain drug sales data, drug attributes, and drug usage for at least one drug in a historical time period.

[0036] Among them, drug sales data may refer to data related to drug sales. Drug sales data may include data such as the sales quantity, sales time and sales location of drugs. Drug sales data can be obtained through channels such as offline platforms and online platforms. Drug sales data is used to provide detailed information on drug sales at different times and locations, providing a data basis for predicting the spread of epidemics.

[0037] Among them, the drug attributes may refer to the attributes of which diseases the drug can treat. Drug attributes may include attributes such as the drug name and the type of disease treated. Drug attributes are used to determine the association between the drug and the type of disease treated. For example, the drug attribute of drug A may be that A can treat the common cold and fever; the drug attribute of drug B may be that B can treat bacterial infections and gastrointestinal infections; and the drug attribute of drug C may be that C can treat invasive fungal infections and deep fungal infections.

[0038] Among them, the intended use of a drug may refer to the specific disease that the patient who purchased the drug is expected to treat with the drug. When a patient purchases a drug on an offline platform, the purpose of purchase can be provided to obtain the user's purchase of a certain drug and the intended use of the drug; when a patient purchases a drug on an online drug sales platform, search for a certain disease in the search bar, and the search results include drugs for treating the type of disease. The drug purchased by the user provided by the user is obtained, and the disease is associated with the drug, thereby determining that the intended use of the drug is the disease. The intended use of a drug is used to record the relationship between drug sales data and the intended use of a drug. Exemplarily, when a patient searches for gastrointestinal infection in the search bar and selects amoxicillin under the gastrointestinal infection menu, the intended use of the drug is to treat gastrointestinal infection.

[0039] S102. Predict the sales data of each drug in a future time period based on the drug sales data of each drug in a historical time period.

[0040] Among them, predicting drug sales data can help users predict the spread of diseases and prevent and control possible epidemics. For example, suppose that the sales data of drug A in a certain area has shown a clear upward trend in the past few months, and the sales volume is associated with the incidence of a certain respiratory disease. Based on the drug sales data, the sales volume of the drug in the future can be predicted, and the possible occurrence of epidemics in the area can be inferred.

[0041] Among them, the future time period may refer to the future forecast time range of drug sales data. The future time period is used to determine the time range to which the predicted drug sales data belongs. Exemplarily, when the future time period is from the 10th to the 20th of a future month, the historical time period may refer to the 10th to the 20th of the previous month adjacent to the future month; when the future time period is from October to December of a future year, the historical time period may refer to October to December of the previous year adjacent to the future year.

[0042] In a specific example, assuming that in a certain region, the drug sales data of the historical time period show that the sales volume of drug A is higher in winter every year, the drug sales data of drug A in the future time period (ie, the upcoming winter) can be predicted.

[0043] S103. Determine the epidemiological orientation relationship of each drug based on drug sales data, drug properties, and drug usage of each drug.

[0044] Among them, the epidemic orientation relationship can refer to the frequency and distribution of use of a certain drug for different epidemics during use. The treatment orientation of drugs for various epidemics can be determined based on the drug sales data, drug properties and drug use purposes of each type of drug. In a specific example, among the drug sales data, drug properties and drug use purposes of each drug, the drug sales data, drug properties and drug use purposes of the same type of drug are counted. For example, the drug sales data, drug properties and drug use purposes of all drug A are obtained. There are 100 drug sales data of all drug A, and the drug properties of drug A can treat epidemics of types a, b and c. Among them, the drug sales data of drug A for the treatment of epidemic a is 50, the drug sales data of drug A for the treatment of epidemic b is 30, and the drug sales data of drug A for the treatment of epidemic c is 20, that is, the epidemic orientation relationship of drug A, and the epidemic orientation relationship of drug A can be expressed by the following formula:

[0045]

[0046] That is, the epidemic pointing relationship can be expressed by the following formula:

[0047] YX =k1x1+k2x2+…+k n x n

[0048] Where Y x represents the epidemic pointing relationship of drug X; n represents the number of epidemics that can be treated by drug X; x n represents the nth epidemic; k n represents the weight coefficient of the nth epidemic, where k1 to k n The cumulative sum is 1.

[0049] S104. Determine the spatiotemporal information of the epidemics pointed to by each drug based on the epidemic pointing relationship of each drug and the drug sales data of each drug in a future time period.

[0050] Among them, the spatiotemporal information of the spread of an epidemic may refer to the fusion characteristic information of the temporal characteristic information and spatial characteristic information of the epidemic. The spatiotemporal information of the spread of an epidemic may include the spread of the epidemic in different time periods and different locations. Among them, the spread of a disease usually has seasonal, periodic or sudden characteristics. The temporal characteristic information of an epidemic can extract the peak period of the epidemic. For example, influenza is usually prevalent in autumn and winter, and the seasonal characteristics can be analyzed as a temporal feature; the spatial characteristic information of an epidemic may refer to the geographical distribution of the epidemic. For example, some epidemics may have different transmission patterns in a city, a village or a specific area. The spatial characteristic information is analyzed through the sales data of drugs and the geographical distribution data of the disease to reveal the hot spots of the spread of the epidemic. The spatiotemporal information of the spread of an epidemic can combine time and space data for analysis, which can more accurately predict the spread trend of the disease and provide predictions of the spread of the epidemic and suggestions for resource allocation.

[0051] It can be seen that in the embodiments of the present application, by obtaining drug sales data, drug properties and drug usage in historical time periods, a comprehensive understanding of drugs can be achieved, thereby providing a data basis for subsequent predictions; by analyzing historical drug sales data, it is possible to predict future drug sales trends and provide a reference for epidemic predictions; by combining drug sales data, properties and uses, it is possible to analyze the potential relationship between drugs and epidemics; by combining the directional relationship between drugs and epidemics and future sales data, it is possible to predict the spatiotemporal characteristics of epidemic spread, thereby improving the accuracy of epidemic predictions.

[0052] In an optional embodiment, "determining the epidemiological pointing relationship of each drug based on its sales data, drug properties and drug usage" is refined into "determining the disease type corresponding to each drug based on its properties; for each drug, determining the pointing probability of the drug for each corresponding disease type based on the disease type, drug usage and drug sales data corresponding to the drug; the epidemiological pointing relationship of the drug includes the pointing probability of the drug for each corresponding disease type" to improve the operation of epidemic prediction.

[0053] It should be noted that for the parts not described in detail in the embodiments of the present invention, reference may be made to the descriptions of other embodiments. Figure 2 A flowchart of an epidemic prediction method provided by an embodiment of the present invention.

[0054] See also Figure 2 The epidemic prediction methods shown include:

[0055] S201. Obtain drug sales data, drug attributes, and drug usage for at least one drug in a historical time period.

[0056] S202. Predict the drug sales data of each drug in a future time period based on the drug sales data of each drug in a historical time period.

[0057] S203. Determine the disease type corresponding to each drug based on the drug properties of each drug.

[0058] Specifically, according to the drug properties of each drug, it is possible to obtain which diseases each drug can treat, and then determine the disease type corresponding to each drug. In a specific example, according to the drug properties of drug A, it can be determined that drug A corresponds to disease types a, b, and c; according to the drug properties of drug B, it can be determined that drug B corresponds to disease types d, e, and f.

[0059] S204. For each drug, determine the pointing probability of the drug for each corresponding disease type according to the disease type, drug usage and drug sales data; the epidemiological pointing relationship of the drug includes the pointing probability of the drug for each corresponding disease type.

[0060] Among them, the pointing probability may refer to the probability value of a certain drug being used to treat a certain disease. The pointing probability can be determined by the disease type, drug usage and drug sales data corresponding to the drug. The pointing probability reflects the specific usage trend in the drug sales data of the historical time period. In a specific example, there are 100 drug sales data of all drug A, and the drug attribute of drug A can treat epidemics of types a, b and c. Among them, there are 50 drug sales data of drug A for the treatment of epidemic a, 30 drug sales data of drug A for the treatment of epidemic b, and 20 drug sales data of drug A for the treatment of epidemic c. The pointing probability of drug A corresponding to type a epidemic is 50%, the pointing probability of drug A corresponding to type b epidemic is 30%, and the pointing probability of drug A corresponding to type c epidemic is 20%.

[0061] S205. Determine the spatiotemporal information of the epidemics pointed to by each drug based on the epidemiological orientation relationship of each drug and the drug sales data of each drug in a future time period.

[0062] It can be seen that in this embodiment, the attributes of the drug can help determine the type of disease that the drug treats, and can simplify the classification of drugs in disease treatment; the drug's disease type, usage and sales data can be used to determine the probability of the drug pointing to a specific disease type; based on the drug's sales data, usage and disease type, the distribution of the drug in actual use can be more objectively analyzed; through the pointing probability, the relationship between the drug and the disease type can be quantified, providing data support for subsequent analysis and decision-making.

[0063] In some embodiments, according to the disease type corresponding to the drug, the drug usage and drug sales data, determining the pointing probability of the drug for each corresponding disease type includes:

[0064] Based on the drug sales data of the drug, the sales quantity corresponding to each drug use of the drug is counted;

[0065] Determine the disease type corresponding to each drug use according to the disease type corresponding to the drug and the disease types included in the drug use of the drug;

[0066] Based on the disease types corresponding to the usage of each drug and the sales volume corresponding to the usage of each drug, the probability of the drug being targeted at each corresponding disease type is calculated.

[0067] Specifically, based on the sales data of all drugs, the sales quantities corresponding to the various drug uses are counted, so that the sales volume of each drug under different drug uses can be clarified, that is, the sales quantity of each drug for treating which disease can be determined.

[0068] The disease type corresponding to a drug may refer to the disease type that the drug itself can treat, for example, a certain antibiotic can be used to treat bacterial infection. The disease type included in the drug usage of a drug may refer to the disease type that the drug is actually used to treat after it is purchased.

[0069] Based on the relationship between the drug's intended use and the disease type, as well as the sales volume, the probability of the drug being targeted at a certain disease type is calculated. The sales volume corresponding to each drug's intended use may refer to the sales volume of the drug under different drug intended uses. For example, in all drug sales records, the sales volume for a certain disease accounts for the proportion of the total sales volume. For example, if the total sales volume of a drug is 1,000 pieces, of which 500 pieces are used to treat disease A, the probability of targeting disease A is 50%.

[0070] It can be seen that in this embodiment, by statistically analyzing the correspondence between drug sales data and its intended use, the sales performance of drugs in the treatment of different diseases can be accurately grasped, providing data support for subsequent predictions and analysis; by clarifying the relationship between drug use and disease types, the role of drugs in the treatment of different diseases can be clearly defined, which helps to provide an effective basis for future drug sales data forecasts; by calculating the pointing probability of drugs under different disease types, the potential sales trend of drugs in the treatment of various diseases can be quantified, thereby improving the accuracy of drug sales forecasts.

[0071] In some embodiments, obtaining drug sales data, drug attributes, and drug usage of at least one drug in a historical time period includes:

[0072] Acquire medicine purchase information provided by at least one user, where the medicine purchase information includes online medicine purchase information and / or offline medicine purchase information;

[0073] Collect statistics on the purchase information of each medicine and obtain the sales data of each medicine;

[0074] According to the drug identification in each medicine purchase information, query the drug attributes of each drug;

[0075] Extract drug usage from medical purchase information.

[0076] The medicine purchase information may refer to all relevant data collected during the medicine purchase process. The medicine purchase information may be obtained from online medicine purchase information and / or offline medicine purchase information. The medicine purchase information may include basic sales information of the medicine, such as purchase time, purchase quantity and purchase location, and specific information of the purchaser, such as the type of disease the purchaser purchases the medicine for treatment.

[0077] Among them, the identification of a drug may refer to a mark or number used to uniquely identify a drug. The identification of a drug may include information such as the manufacturer of the drug, the type of drug, the dosage form, the specification, the batch number, etc. The purpose of drug identification is to ensure that each drug can be accurately identified, to facilitate the tracing of the production and circulation chain of the drug, and to avoid confusion in sales, use, monitoring and management. For example, in a certain medical purchase information, the identification of drug A may be ABC123-XXXX, which indicates that drug ABC is a batch produced by a certain manufacturer in the year XXXX, and the name of the drug is 123.

[0078] The extraction of the purpose of use of the drug from the drug purchase information may be to use the type of disease for which the purchaser purchases the drug to treat as the purpose of use of the drug in the drug purchase information.

[0079] It can be seen that in this embodiment, by obtaining the medical purchase information provided by the user, the data on drug sales on various platforms can be comprehensively collected. By including online and offline purchase information, the drug sales situation of different channels can be obtained, providing a multiple data source for predicting the future sales trend of drugs; by statistics on medical purchase information, the detailed data on drug sales can be effectively summarized, such as sales quantity, sales time and sales location, etc.; by statistical sales data, accurate drug sales data can be provided for subsequent analysis and prediction, thereby better predicting future changes in drug demand; by querying the properties of drugs through drug identification, the specific information of each drug can be accurately identified, such as ingredients, uses and manufacturers, etc.; by extracting the use of drugs, the treatment direction or indications of drugs can be further clarified, providing support for the association analysis between drugs and specific diseases.

[0080] In some embodiments, based on the epidemic direction relationship of each drug and the drug sales data of each drug in a future time period, the spatiotemporal information of the epidemic spread pointed by each drug is determined, including:

[0081] Determine the predicted purchase time, predicted purchase address and predicted purchase quantity of each drug based on the drug sales data of each drug in the future time period;

[0082] According to the pointing probability of each drug to each corresponding disease type and the predicted purchase time, predicted purchase address and predicted purchase quantity of each drug, the distribution time, distribution address and distribution quantity of each disease type are determined; the propagation spatiotemporal information includes the distribution time, distribution address and distribution quantity.

[0083] The predicted purchase time may refer to the predicted purchase time of a drug in a certain time period in the future. For example, assuming that the sales volume of a drug increases significantly in winter every year, then based on the historical sales data and the epidemiological relationship, it can be predicted that the drug will reach a peak in purchase in a certain time period in the next winter.

[0084] Among them, predicting the purchase address can refer to predicting the geographical location of drug purchase in the future time period, that is, in which regions patients are most likely to purchase a certain drug. By analyzing the purchase address in the historical sales data of the drug, combined with the epidemiological patterns, disease types and drug usage in different regions, it is possible to predict the future demand hotspots for the drug. For example, assuming that a certain drug has a high sales volume in a certain city in the past few months, and these cities happen to be areas where a certain type of disease is prevalent, based on this information, it can be predicted that the drug will be concentrated in these cities in the future time period.

[0085] The predicted purchase quantity may refer to predicting the expected purchase quantity of a drug in a future time period. For example, assuming that the sales volume of a drug in the past winter months was XXX, and the drug is mainly used to treat common winter diseases, it can be predicted that the sales volume of the drug in the future winter will also be XXX.

[0086] The distribution time may refer to the distribution of the demand for the drug at each time node in the future time period based on the predicted purchase quantity and purchase time.

[0087] The distribution address may refer to specific geographical locations or regions where the sales of drugs will be distributed in a certain period of time in the future based on predicted purchase addresses and purchase quantities.

[0088] The distribution quantity can refer to the distribution of sales quantity or demand quantity in a specific geographical location within a certain period of time in the future.

[0089] It can be seen that in this embodiment, by analyzing the drug sales data in the future time period, the purchase time, purchase address and purchase quantity of the drug can be accurately predicted; by combining the probability information of the drug pointing to the disease type and the predicted purchase data, the spatiotemporal information of the spread of the disease type can be determined, which can reveal the distribution characteristics of the disease type in the future.

[0090] In some embodiments, based on the drug sales data of each drug in a historical time period, predicting the drug sales data of each drug in a future time period includes:

[0091] Determine first sales data and second sales data according to the drug sales data of each drug in a historical time period, wherein the first sales data includes the sales time and sales quantity of each drug, and the second sales data includes the sales address and sales quantity of each drug;

[0092] Determining a time feature based on the first sales data;

[0093] Determining spatial characteristics based on the second sales data;

[0094] According to the time characteristics and spatial characteristics, the drug sales data of each drug in the future time period is determined.

[0095] The first sales data may refer to time information about drug sales, and the first sales data may include the sales time and sales quantity of the drug. By analyzing the sales time and quantity, it is possible to determine the time pattern of drug sales and its changing trend.

[0096] Among them, the time feature may refer to the pattern and regularity of drug sales time extracted from the first sales data, and the time feature reflects the characteristics of drug sales changing over time.

[0097] In a specific example, the first sales data is divided into five categories of sales data according to hourly sales, daily sales, weekly sales, monthly sales and quarterly sales; and five multi-head self-attention expert networks with the same structure are used to output five time features at different granularities. The network structure diagram of the multi-head self-attention expert network is shown in Figure 3 As shown, the time-sales data may refer to the first sales data, and the hourly sales data, daily sales data, weekly sales data, monthly sales data and quarterly sales data are respectively input into five multi-head self-attention expert networks with the same structure, and after a gated recurrent unit, five time features at different granularities are output.

[0098] The second sales data may include information related to the geographical location of drug sales, and may include the sales address and sales quantity of the drug. By analyzing the sales address and sales quantity, the demand distribution of the drug in different geographical locations can be determined, revealing regional sales characteristics.

[0099] The spatial feature may refer to information related to the geographical location of drug sales extracted from the second sales data, and the spatial feature reflects the sales distribution of drugs in different geographical areas.

[0100] In a specific example, based on the time characteristics and spatial characteristics, the drug sales data of each drug in the future time period can be determined by fusing the time characteristics and spatial characteristics. In order to effectively fuse the time characteristics and spatial characteristics determined above, the time weight and spatial feature weight can be adaptively adjusted according to the input data of different drug names. The MoE module with a gating mechanism can capture the dependency between different time granularities and spatial area granularities, and dynamically adjust the focus of the model, thereby reducing the overfitting of the model to a specific time granularity or a specific geographical area, and improving the robustness and prediction accuracy of the model. The formula for operating the gating mechanism is as follows:

[0101]

[0102] In the formula, H s,t Represents the feature embedding vector after encoding by the gating mechanism; (W s,t , W i,s , W j,t , b s,t ) represents the trainable parameters; σ represents the sigmoid function, which is used to control the flow of information; E i,s represents the spatial eigenvector, E j,t represents the time feature vector; Represents the temporal characteristic noise vector; Represents the spatial characteristic noise vector.

[0103] It can be seen that in this embodiment, by distinguishing the time dimension and space dimension of sales data, the sales situation of drugs can be analyzed in more detail; by refining the structure of sales data, the impact of sales time and sales location on sales volume can be better captured; by analyzing sales time data, the cyclical fluctuations or long-term trends of drug sales can be identified; by analyzing sales addresses and sales quantities, the differences in drug demand in different regions can be identified; and by integrating time and space characteristics, a more comprehensive sales forecast can be provided.

[0104] In some embodiments, determining the spatial feature according to the second sales data includes:

[0105] Generate a spatial matrix according to the second sales data, wherein the matrix position of an element in the spatial matrix is ​​the sales address of the drug corresponding to the element, and the value of the element is the sales quantity of the drug corresponding to the element;

[0106] Generate spatial features based on the spatial matrix.

[0107] Among them, the spatial matrix can refer to a matrix that combines the sales address and sales quantity of a drug. The matrix position of each element in the matrix corresponds to a specific sales address, and the positional relationship between different elements in the matrix corresponds to the positional relationship between the sales addresses of the corresponding drugs. The value of each matrix element represents the sales quantity at the sales location. Exemplarily, assuming that the sales quantity of drug A in the four adjacent cities A, B, C and D is 100, 200, 300 and 400 respectively, the spatial matrix is ​​(100, 200, 300, 400).

[0108] In a specific example, the process of generating spatial features according to the spatial matrix is ​​as follows, wherein the spatial features may include global spatial features and local spatial features. Figure 4A schematic diagram of the network structure of a multi-scale feature pyramid provided by an embodiment of the present invention. The original spatial matrix, the 1 / 4-sized spatial matrix and the 4-sized spatial matrix are used as inputs to the multi-scale feature pyramid; the 1 / 4-sized spatial matrix passes through 3 layers of residual convolutional networks to obtain a 1 / 32-sized feature mapping matrix; the original spatial matrix passes through 3 convolutional layers to obtain a 1 / 4-sized feature mapping matrix; the 4-sized original spatial matrix passes through 3 layers of residual convolutional networks and 3 convolutional layers separately to obtain 1-sized and 2-sized feature mapping matrices respectively, and finally all feature mapping matrices pass through a gated network to output global spatial features and local spatial features.

[0109] It can be seen that in this embodiment, by generating a spatial matrix based on the second sales data, the spatial distribution of drug sales can be quantified and visualized, and the sales of drugs in different regions can be intuitively displayed, thereby helping to analyze sales hotspots, demand distribution, etc.; by generating a spatial matrix, spatial features related to the sales address can be extracted therefrom; by generating spatial features based on the spatial matrix, the spatial behavior pattern of drug sales can be modeled; by constructing and analyzing the spatial matrix, the accuracy of epidemic prediction can be improved.

[0110] Figure 5 A schematic diagram of the structure of an epidemic prediction device provided by an embodiment of the present invention. The embodiment of the present invention is applicable to the situation of epidemic prediction, the device can execute the epidemic prediction method, and the device can be implemented in the form of hardware and / or software.

[0111] See also Figure 5 The epidemic prediction device shown includes: a data acquisition module 501, a prediction module 502, a relationship determination module 503 and an information determination module 504, wherein:

[0112] The data acquisition module 501 is used to acquire drug sales data, drug attributes and drug usage of at least one drug in a historical time period;

[0113] Prediction module 502, used to predict the drug sales data of each drug in a future time period based on the drug sales data of each drug in a historical time period;

[0114] The relationship determination module 503 is used to determine the epidemiological orientation relationship of each drug according to the drug sales data, drug attributes and drug usage of each drug;

[0115] The information determination module 504 is used to determine the spatiotemporal information of the epidemics pointed to by each drug based on the epidemic pointing relationship of each drug and the drug sales data of each drug in the future time period.

[0116] The technical solution of the embodiment of the present invention can achieve a comprehensive understanding of drugs by acquiring drug sales data, drug properties and drug usage in historical time periods, thereby providing a data basis for subsequent predictions; by analyzing historical drug sales data, it is possible to predict future drug sales trends and provide a reference for epidemic predictions; by combining drug sales data, properties and uses, it is possible to analyze the potential relationship between drugs and epidemics; by combining the directional relationship between drugs and epidemics and future sales data, it is possible to predict the spatiotemporal characteristics of epidemic transmission, thereby improving the accuracy of epidemic predictions.

[0117] In some embodiments, in determining the epidemiological relationship of each drug according to the drug sales data, drug attributes and drug usage of each drug, the relationship determination module 503 is specifically used to:

[0118] According to the drug properties of each drug, determine the disease type corresponding to each drug;

[0119] For each drug, the pointing probability of the drug for each corresponding disease type is determined based on the drug's corresponding disease type, drug usage and drug sales data; the drug's epidemiological pointing relationship includes the pointing probability of the drug for each corresponding disease type.

[0120] In some embodiments, in determining the pointing probability of a drug for each corresponding disease type according to the disease type corresponding to the drug, the drug usage and the drug sales data, the relationship determination module 503 is specifically used to:

[0121] Based on the drug sales data of the drug, the sales quantity corresponding to each drug use of the drug is counted;

[0122] Determine the disease type corresponding to each drug use according to the disease type corresponding to the drug and the disease types included in the drug use of the drug;

[0123] Based on the disease types corresponding to the usage of each drug and the sales volume corresponding to the usage of each drug, the probability of the drug being targeted at each corresponding disease type is calculated.

[0124] In some embodiments, in terms of obtaining drug sales data, drug attributes, and drug usage of at least one drug in a historical time period, the data acquisition module 501 is specifically used to:

[0125] Acquire medicine purchase information provided by at least one user, where the medicine purchase information includes online medicine purchase information and / or offline medicine purchase information;

[0126] Collect statistics on the purchase information of each medicine and obtain the sales data of each medicine;

[0127] According to the drug identification in each medicine purchase information, query the drug attributes of each drug;

[0128] Extract drug usage from medical purchase information.

[0129] In some embodiments, in determining the spatiotemporal information of the epidemics pointed to by each drug based on the epidemic pointing relationship of each drug and the drug sales data of each drug in the future time period, the information determination module 504 is specifically used to:

[0130] Determine the predicted purchase time, predicted purchase address and predicted purchase quantity of each drug based on the drug sales data of each drug in the future time period;

[0131] According to the pointing probability of each drug to each corresponding disease type and the predicted purchase time, predicted purchase address and predicted purchase quantity of each drug, the distribution time, distribution address and distribution quantity of each disease type are determined; the propagation spatiotemporal information includes the distribution time, distribution address and distribution quantity.

[0132] In some embodiments, in predicting the drug sales data of each drug in a future time period based on the drug sales data of each drug in a historical time period, the prediction module 502 is specifically used to:

[0133] Determine first sales data and second sales data according to the drug sales data of each drug in a historical time period, wherein the first sales data includes the sales time and sales quantity of each drug, and the second sales data includes the sales address and sales quantity of each drug;

[0134] Determining a time feature based on the first sales data;

[0135] Determining spatial characteristics based on the second sales data;

[0136] According to the time characteristics and spatial characteristics, the drug sales data of each drug in the future time period is determined.

[0137] In some embodiments, in determining the spatial characteristics according to the second sales data, the prediction module 502 is specifically configured to:

[0138] Generate a spatial matrix according to the second sales data, wherein the matrix position of an element in the spatial matrix is ​​the sales address of the drug corresponding to the element, and the value of the element is the sales quantity of the drug corresponding to the element;

[0139] Generate spatial features based on the spatial matrix.

[0140] The epidemic prediction device provided in the embodiment of the present invention can execute the epidemic prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the epidemic prediction method.

[0141] Figure 6 A schematic diagram of the structure of an epidemic prediction device provided by an embodiment of the present invention.

[0142] like Figure 6 As shown, the epidemic prediction device 600 includes at least one processor 601, and a memory connected to the at least one processor 601, such as a read-only memory (ROM) 602, a random access memory (RAM) 603, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 601 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 602 or the computer program loaded from the storage unit 608 to the random access memory (RAM) 603. In RAM603, various programs and data required for the operation of the epidemic prediction device 600 can also be stored. The processor 601, ROM602 and RAM603 are connected to each other via a bus 604. An input / output (I / O) interface 608 is also connected to the bus 604.

[0143] A number of components in the epidemic prediction device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the epidemic prediction device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0144] Processor 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of processor 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. Processor 601 performs the various methods and processes described above, such as an epidemic prediction method.

[0145] In some embodiments, the epidemic prediction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the epidemic prediction device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the processor 601, one or more steps of the epidemic prediction method described above may be performed. Alternatively, in other embodiments, the processor 601 may be configured to perform the epidemic prediction method in any other appropriate manner (e.g., by means of firmware).

[0146] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0147] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0148] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer 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 foregoing.

[0149] To provide interaction with a user, the systems and techniques described herein may be implemented on an operation detection device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which a user can provide input to the epidemic prediction device. Other types of devices may also be used to provide interaction with a user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0150] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0151] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS (Virtual Private Server) services.

[0152] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0153] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting an epidemic, characterized in that: The method comprises: Obtain drug sales data, drug attributes, and drug usage for at least one drug in a historical time period; Based on the drug sales data of each of the drugs in the historical time period, predict the drug sales data of each of the drugs in the future time period; Determine the epidemiological relationship of each of the drugs based on drug sales data, drug properties and drug usage of each of the drugs; According to the epidemiological orientation relationship of each of the drugs and the drug sales data of each of the drugs in a future time period, the spatiotemporal information of the spread of the epidemic pointed to by each of the drugs is determined.

2. The method according to claim 1, characterized in that: Determining the epidemiological orientation relationship of each of the drugs according to the drug sales data, drug properties and drug usage of each of the drugs includes: Determining the disease type corresponding to each of the drugs according to the drug properties of each of the drugs; For each of the drugs, the pointing probability of the drug for each corresponding disease type is determined based on the disease type, drug usage and drug sales data corresponding to the drug; the epidemiological pointing relationship of the drug includes the pointing probability of the drug for each corresponding disease type.

3. The method according to claim 2, characterized in that The determining, based on the disease type corresponding to the drug, the drug usage and drug sales data, the pointing probability of the drug for each corresponding disease type includes: According to the drug sales data of the drug, the sales quantity corresponding to each drug use of the drug is counted; Determine the disease type corresponding to each use of the drug according to the disease type corresponding to the drug and the disease types included in the drug use of the drug; According to the disease types corresponding to the uses of the drugs and the sales quantities corresponding to the uses of the drugs, the targeting probabilities of the drugs for the corresponding disease types are calculated.

4. The method according to claim 1, characterized in that: The step of obtaining drug sales data, drug attributes, and drug usage of at least one drug in a historical time period includes: Acquire medicine purchase information provided by at least one user, wherein the medicine purchase information includes online medicine purchase information and / or offline medicine purchase information; Collecting statistics on the purchase information of each medicine to obtain the sales data of each medicine; According to the identification of the medicine in each of the medicine purchase information, query the medicine attribute of each of the medicines; The purpose of drug use is extracted from the drug purchasing information.

5. The method according to claim 3, characterized in that: Determining the spatiotemporal information of the epidemics pointed to by each of the drugs based on the epidemic pointing relationship of each of the drugs and the drug sales data of each of the drugs in a future time period includes: Determine the predicted purchase time, predicted purchase address and predicted purchase quantity of each of the drugs based on the drug sales data of each of the drugs in the future time period; According to the pointing probability of each of the drugs for each of the corresponding disease types and the predicted purchase time, predicted purchase address and predicted purchase quantity of each of the drugs, the distribution time, distribution address and distribution quantity of each of the disease types are determined; the propagation spatiotemporal information includes the distribution time, distribution address and distribution quantity.

6. The method according to claim 1, characterized in that The method of predicting the drug sales data of each drug in a future time period based on the drug sales data of each drug in a historical time period includes: Determine first sales data and second sales data according to the drug sales data of each of the drugs in a historical time period, wherein the first sales data includes the sales time and sales quantity of each of the drugs, and the second sales data includes the sales address and sales quantity of each of the drugs; Determining a time feature based on the first sales data; determining spatial characteristics according to the second sales data; The drug sales data of each of the drugs in a future time period is determined according to the time characteristics and the space characteristics.

7. The method according to claim 6, characterized in that The determining of the spatial characteristics according to the second sales data includes: Generate a space matrix according to the second sales data, wherein the matrix position of an element in the space matrix is ​​the sales address of the drug corresponding to the element, and the value of the element is the sales quantity of the drug corresponding to the element; A spatial feature is generated according to the spatial matrix.

8. An epidemic prediction device, characterized in that: include: A data acquisition module, used to acquire drug sales data, drug attributes and drug usage of at least one drug in a historical time period; A prediction module, used to predict the drug sales data of each of the drugs in a future time period based on the drug sales data of each of the drugs in a historical time period; A relationship determination module, for determining the epidemiological orientation relationship of each of the drugs according to the drug sales data, drug attributes and drug usage of each of the drugs; The information determination module is used to determine the spatiotemporal information of the epidemic spread pointed by each of the drugs based on the epidemic pointing relationship of each of the drugs and the drug sales data of each of the drugs in a future time period.

9. An epidemic prediction device, characterized in that: The epidemic prediction device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the epidemic prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the epidemic prediction method according to any one of claims 1 to 7 when executed.

Images