Data processing method and device

By extracting numerical and timing features from supply chain node data and analyzing them in combination with preset models, the problem of feature correlation in the prior art is solved, and the accuracy of supply chain data analysis and the user experience of management end are improved.

CN119090112APending Publication Date: 2024-12-06BEIJING JINGDONG QIANSHITECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310666978.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-06-06
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The prior art fails to effectively consider the correlation between relatively independent features in node data analysis in the supply chain, resulting in low analysis accuracy and difficulty in improving supply chain management.

Method used

By extracting numerical and timing characteristics from the supply chain node data, inputting the trained tree classification model and timing model respectively, predicting the order change trend, and generating data analysis results based on the results, and providing them to the management end.

Benefits of technology

It improves the accuracy of supply chain node data analysis, improves the user experience of the management end, and can more effectively improve supply chain management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119090112A_ABST
    Figure CN119090112A_ABST
Patent Text Reader

Abstract

The invention discloses a data processing method and device, and relates to the technical field of intelligent supply chains. A specific embodiment of the method comprises the following steps: extracting numerical value type features and time sequence type features from node data of a supply chain, and respectively inputting the features into a trained preset tree classification model and a trained preset time sequence model so as to respectively predict an order change trend of the node, determining a data analysis result of the node by combining the obtained order change trend, and sending the data analysis result to a management end of the node; according to the embodiment of the invention, different types of characteristics with relevance are extracted from the data, and data analysis is carried out in combination with prediction results of different prediction models, so that the accuracy of data analysis of the supply chain nodes is improved to a greater extent, and the user experience of a management end for managing the supply chain is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of smart supply chain technology, and in particular to a data processing method and device. Background Art

[0002] With the application of artificial intelligence in the supply chain, more and more neural network models are used to analyze the data generated by nodes in the supply chain to guide the nodes to make supply adjustments.

[0003] At present, relatively independent features contained in the historical data of nodes, such as order volume, complaint information, etc., are usually used to train a neural network model, and then the trained neural network model is used to analyze the supply situation of the node. In the existing method, the accuracy of the data analysis generated by the nodes in the supply chain is low because the correlation between relatively independent features is not considered and the data is only analyzed through one neural network model, which makes it difficult to effectively improve the entire supply chain. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a data processing method and device, which can extract numerical features and time series features from node data of a supply chain, and input the features into a trained preset tree classification model and a trained preset time series model respectively, so as to respectively predict the order change trend of the node, and determine the data analysis result of the node in combination with the obtained order change trend and send it to the management end of the node; the embodiment of the present invention extracts different types of correlated features from the data, and performs data analysis in combination with the prediction results of different prediction models, thereby greatly improving the accuracy of data analysis of supply chain nodes and improving the user experience of the management end of managing the supply chain.

[0005] To achieve the above-mentioned purpose, according to one aspect of an embodiment of the present invention, a data processing method is provided, which is characterized in that it includes: obtaining node data generated by any node on the supply chain within a set time range; extracting a numerical feature matrix including multiple numerical features from the node data, inputting the numerical feature matrix including multiple numerical features into a preset tree classification model, and using the preset tree classification model to predict the first order change trend of the node; extracting a time series feature matrix including multiple time series features from each of the numerical features, inputting the time series feature matrix including multiple time series features into a preset time series model, and using the preset time series model to predict the second order change trend of the node; generating a data analysis result for the node in combination with the first order change trend and the second order change trend, and providing the data analysis result to the management end corresponding to the node.

[0006] Optionally, extracting a numerical feature matrix including multiple numerical features from the node data includes: acquiring data corresponding to a first numerical feature corresponding to one or more unit times from the node data; calculating data corresponding to one or more second numerical features of extended unit times based on the data corresponding to the first numerical feature; and constructing a numerical feature matrix using the data corresponding to the first numerical feature and the data corresponding to the second numerical feature.

[0007] Optionally, extracting a time series feature matrix including multiple time series features from each of the numerical features includes: constructing multiple unit times of time continuity and multiple extended unit times in time order; extracting time series features corresponding to multiple unit times of time continuity from the data corresponding to the first numerical feature; extracting time series features corresponding to multiple extended unit times of time continuity from the data corresponding to the second numerical feature.

[0008] Optionally, the data processing method further includes: obtaining historical node data of multiple nodes included in the supply chain; extracting a variety of historical numerical features from the historical node data to iteratively train a tree classification model, and determining the trained tree classification model as a preset tree classification model; extracting a variety of time series features from the multiple historical numerical features to iteratively train a time series model, and determining the trained time series model as a preset time series model.

[0009] Optionally, after acquiring the historical node data of the multiple nodes included in the supply chain, it also includes: acquiring multiple node levels divided for the multiple nodes included in the supply chain; for each of the node levels, screening out target historical node data that meets the node level from the historical node data of the multiple nodes; using the historical node data to extract multiple numerical features to iteratively train the tree classification model, including: using the target historical node data of the multiple node levels to extract multiple historical numerical features to iteratively train the tree classification model; using the multiple numerical features extracted from the historical node data to extract multiple time series features to iteratively train the time series model, including: using the multiple historical numerical features to extract multiple time series features to iteratively train the time series model.

[0010] Optionally, the data processing method further includes: combining the first order change trend and the first second order change trend to generate a data analysis result for the node, including: using the first order change trend and the second order change trend to predict an order increase or order loss at the node, and generating a data analysis result indicating an order increase or order loss.

[0011] Optionally, after selecting the target historical node data satisfying the node level from the historical node data of the plurality of nodes, the method further includes: when taking order increase as the training target of model training, increasing the proportion of historical numerical features indicating order increase extracted from the target historical node data, and increasing the proportion of time series features indicating order increase extracted from the historical numerical features.

[0012] To achieve the above-mentioned purpose, according to a second aspect of an embodiment of the present invention, a data processing device is provided, characterized in that it comprises: a data acquisition module, an execution prediction module and a data analysis module; wherein:

[0013] The data acquisition module is used to acquire node data generated by any node in the supply chain within a set time range;

[0014] The execution prediction module is used to extract a numerical feature matrix including multiple numerical features from the node data, input the numerical feature matrix including multiple numerical features into a preset tree classification model, and use the preset tree classification model to predict the first order change trend of the node; extract a time series feature matrix including multiple time series features from each of the numerical features, input the time series feature matrix including multiple time series features into a preset time series model, and use the preset time series model to predict the second order change trend of the node;

[0015] The data analysis module is used to generate a data analysis result for the node in combination with the first order change trend and the second order change trend, and provide the data analysis result to the management end corresponding to the node.

[0016] Optionally, the data processing device is used to extract a numerical feature matrix including multiple numerical features from the node data, including: obtaining data corresponding to a first numerical feature corresponding to one or more unit times from the node data; calculating data corresponding to a second numerical feature of one or more extended unit times based on the data corresponding to the first numerical feature; and constructing a numerical feature matrix using the data corresponding to the first numerical feature and the data corresponding to the second numerical feature.

[0017] Optionally, the data processing device is used to extract a time series feature matrix including multiple time series features from each of the numerical features, including: constructing multiple unit times of continuous time and multiple extended unit times in time order; extracting time series features corresponding to multiple unit times of continuous time from the data corresponding to the first numerical feature; extracting time series features corresponding to multiple extended unit times of continuous time from the data corresponding to the second numerical feature.

[0018] Optionally, the data processing device is further used to: obtain historical node data of multiple nodes included in the supply chain; extract a variety of historical numerical features from the historical node data to iteratively train a tree classification model, and determine the trained tree classification model as a preset tree classification model; extract a variety of time series features from the multiple historical numerical features to iteratively train a time series model, and determine the trained time series model as a preset time series model.

[0019] Optionally, the data processing device, after obtaining the historical node data of the multiple nodes included in the supply chain, further includes: obtaining multiple node levels divided for the multiple nodes included in the supply chain; for each of the node levels, screening out target historical node data that meets the node level from the historical node data of the multiple nodes; the iterative training of the tree classification model by extracting multiple numerical features from the historical node data includes: iterative training of the tree classification model by extracting multiple historical numerical features from the target historical node data of the multiple node levels; the iterative training of the time series model by extracting multiple time series features from the multiple numerical features extracted from the historical node data includes: iterative training of the time series model by extracting multiple time series features from the multiple historical numerical features.

[0020] Optionally, the data processing device is further used to generate a data analysis result for the node in combination with the first order change trend and the first second order change trend, including: using the first order change trend and the second order change trend to predict the order increase or order loss of the node, and generating a data analysis result indicating the order increase or order loss.

[0021] Optionally, the data processing device is used to, after screening out target historical node data that meets the node level from the historical node data of multiple nodes, further include: when taking order increase as the training target of model training, increasing the proportion of historical numerical features indicating order increase extracted from the target historical node data, and increasing the proportion of time series features indicating order increase extracted from the historical numerical features.

[0022] To achieve the above-mentioned purpose, according to the third aspect of an embodiment of the present invention, there is provided an electronic device for data processing, characterized in that it includes: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement any method as described in the above-mentioned data processing method.

[0023] To achieve the above-mentioned purpose, according to a fourth aspect of an embodiment of the present invention, a computer-readable medium is provided, on which a computer program is stored, characterized in that when the program is executed by a processor, a method as described in any of the above-mentioned data processing methods is implemented.

[0024] An embodiment of the above invention has the following advantages or beneficial effects: it can extract numerical features and time series features from the node data of the supply chain, and input the features into the trained preset tree classification model and the trained preset time series model respectively, so as to respectively predict the order change trend of the node, and determine the data analysis result of the node in combination with the obtained order change trend and send it to the management end of the node; the embodiment of the present invention extracts different types of correlated features from the data, and performs data analysis in combination with the prediction results of different prediction models, thereby greatly improving the accuracy of data analysis of supply chain nodes and improving the user experience of the management end of managing the supply chain.

[0025] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0027] Figure 1 It is a flowchart of a data processing method provided by an embodiment of the present invention;

[0028] Figure 2 It is a schematic diagram of a model training process provided by an embodiment of the present invention;

[0029] Figure 3 is a structural schematic diagram of a data processing device provided by an embodiment of the present invention;

[0030] Figure 4 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0031] Figure 5 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0033] It should be noted that the collection, collection, updating, analysis, processing, use, transmission, storage and other aspects of user personal information involved in the technical solution of the present invention are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information, network security and national security.

[0034] After collecting the historical data of the nodes, we will de-identify the data through technical means. When displaying the analysis data results, we will use de-identification or anonymization to desensitize the information to protect information security.

[0035] In order to count the usage of our products / services, we will aggregate, analyze and use the technically processed user data and share the processed statistical information with third parties. We will use secure encryption technology and other methods to ensure that the recipient of the information cannot re-identify a specific individual.

[0036] like Figure 1 As shown, an embodiment of the present invention provides a method for data processing, which may include the following steps:

[0037] Step S101: Obtain node data generated by any node in the supply chain within a set time range.

[0038] Specifically, the supply chain may include one or more nodes, such as online suppliers, stores, wholesale providers of goods, etc. For the management side of the supply chain, by analyzing and predicting the data of each node, the efficiency and accuracy of the management nodes can be improved, so as to improve the entire supply chain through management.

[0039] The node data of a node may include the basic node information of the node, order volume, package volume, item warehousing volume, claim volume, complaint volume, logistics company penetration rate, etc.

[0040] Furthermore, the node data generated by any node on the supply chain within a set time range is obtained, wherein the node to be processed (i.e., any node) can be determined using the node identifier, and the set time range can be one month, one quarter, half a year, etc., depending on the application scenario; multiple node data within the set time range that matches the node identifier can be searched and obtained from the data source according to the timestamp, so as to perform subsequent operations based on the node data.

[0041] Step S102: extracting a numerical feature matrix including a plurality of numerical features from the node data, inputting the numerical feature matrix including the plurality of numerical features into a preset tree classification model, and using the preset tree classification model to predict a first order change trend of the node.

[0042] Specifically, a numerical feature matrix including multiple numerical features is extracted from the node data, wherein the numerical feature matrix corresponding to the numerical features can be extracted from multiple dimensions, such as: node dimension, logistics dimension, business dimension, etc. By extracting numerical features from multiple dimensions, the granularity of the features is reduced and the degree of refinement of the features is increased, thereby improving the accuracy of the model output.

[0043] Further, preferably, based on the data expansion within the unit time contained in the acquired node data, new data within multiple extended unit times are calculated; for example: taking the order volume data in the node data as an example; for example, the set time range corresponding to the acquired node data is three months; then based on the order volume of each day of the node within these three months, the order volume for one week, the order volume for two weeks, the order volume for one month, the order volume for three months, etc. can be calculated; wherein, the order volume for one day is the data within the unit time (that is, the data corresponding to the first numerical feature), the order volume for one week, the order volume for two weeks, the order volume for one month, and the order volume for three months; or the incremental amount and the month-on-month incremental amount of the order volume within one week, two weeks, one month, and adjacent time ranges; or the effective number of days and the statistical value of the number of days (for example: mean, maximum, minimum, median, etc.) for placing orders within one week, two weeks, one month, and three months, respectively, or the time interval and the number of losses (loss refers to the number of orders not being placed within the set time) within one week, two weeks, one month, and three months. The multiple numerical data such as information behavior), loss length, loss proportion, etc. are used to calculate the new data in multiple extended unit times (i.e., the data corresponding to the second numerical feature); in this example, one day is the unit time; one week, two weeks, one month, three months, etc. are multiple extended times; similarly, multiple other data other than the order volume in other unit times can be calculated to obtain the new data in multiple extended unit times; further, a numerical feature matrix is ​​constructed based on the data corresponding to the first numerical feature and the data corresponding to the second numerical feature; that is, the numerical feature matrix including multiple numerical features is extracted from the node data, including: obtaining the data corresponding to the first numerical feature corresponding to one or more unit times from the node data; calculating the data corresponding to the second numerical feature of one or more extended unit times based on the data corresponding to the first numerical feature; and constructing a numerical feature matrix using the data corresponding to the first numerical feature and the data corresponding to the second numerical feature.

[0044] Furthermore, a numerical feature matrix including a plurality of the numerical features is input into a preset tree classification model, and the preset tree classification model is used to predict the first order change trend of the node; in one embodiment of the present invention, the preset tree classification model may be a model obtained through iterative training based on the original GBDT (Gradient Boosting Decision Tree) model; thereby, the trained preset tree classification model is used to predict the first order change trend of the node; wherein, the order change trend may be an order increase or an order loss, and the first order change trend may be a probability value of an order increase, and / or a probability value of an order loss.

[0045] Step S103: extracting a time series feature matrix including multiple time series features from each of the numerical features, inputting the time series feature matrix including multiple time series features into a preset time series model, and using the preset time series model to predict the second order change trend of the node.

[0046] Specifically, a time series feature matrix including multiple time series features is extracted from each of the numerical features; wherein the time series features can reflect the continuity and correlation between multiple time ranges; for example, taking the order quantity as the data of the numerical feature as an example, within a time range of three months, in chronological order, based on the order quantity (i.e., data corresponding to the first numerical feature) of each day (i.e., multiple continuous unit times), a time series feature matrix of time series features associated with the order category is extracted; and, in chronological order, based on the constructed multiple continuous extended unit times (e.g., one week, two weeks, one month, two months, three months, etc.) in time, the order quantity statistics (e.g., sum, mean, etc.), and the order increment and month-on-month order increment in adjacent time ranges; or the effective number of days and day statistics (e.g., mean, maximum, minimum, median, etc.) of placing orders within one week, two weeks, one month, and three months. etc.); or the time intervals of placing orders within one week, two weeks, one month, or three months, the number of churns (for example, churn means that there is no order data within a set period of time), the length of churn, the proportion of churn, and other numerical data, as well as the time correlation between them, to extract the time series feature matrix of the time series features associated with the order category; wherein the order quantity statistics, the increment of orders, the month-on-month increment of orders, the length of churn, the proportion of churn, etc. are the data corresponding to the second numerical feature; that is, the time series feature matrix including multiple time series features is extracted from each of the numerical features, including: constructing multiple unit times and multiple extended unit times of continuous time in chronological order; extracting the time series features corresponding to multiple unit times of continuous time from the data corresponding to the first numerical feature; extracting the time series features corresponding to multiple extended unit times of continuous time from the data corresponding to the second numerical feature.

[0047] Furthermore, a time series feature matrix including a plurality of the time series features is input into a preset time series model, and the preset time series model is used to predict the second order change trend of the node; specifically, in an embodiment of the present invention, the preset time series model may be a trained model obtained by iterative training based on an LSTM (Long short-term memory) model; and the preset time series model is used to predict the second order change trend of the node. The order change trend may be an increase in orders or a loss in orders, and the second order change trend may be a probability value of an increase in orders and / or a probability value of a loss in orders.

[0048] Step S104: Generate a data analysis result for the node in combination with the first order change trend and the second order change trend, and provide the data analysis result to the management end corresponding to the node.

[0049] Specifically, based on the first order change trend and the second order change trend described in steps S102-S103, a data analysis result for the node is generated; wherein the data analysis result can be multiple; for example: based on the probability of order increase and / or order loss of the node, a comprehensive analysis is performed to predict the order increase or order loss of the node; for example, when the probability of order increase indicated by the first order change trend and the second order change trend is greater than the first preset probability threshold, the data analysis result of the node is determined to be an order increase; similarly, the data analysis result of the node is determined to be an order loss; further, the data analysis result indicating order increase or order loss is provided to the management end corresponding to the node; the management end can be a client or a server, or it can be a party that triggers the execution of the data processing method for any node. That is, the data analysis result for the node is generated in combination with the first order change trend and the second order change trend, including: using the first order change trend and the second order change trend to predict the order increase or order loss of the node, and generating a data analysis result indicating order increase or order loss.

[0050] From the description of step S101 to step S104, it can be seen that the embodiments of the present invention extract different types of feature matrices from the node data of the supply chain, including numerical feature matrices and time series feature matrices, and input different feature matrices into different models, respectively, predict the node order change trends, and determine the data analysis results for the nodes (order increase or order loss, etc.) in combination with different order change trends; by calculating and extracting numerical features and time series features, the time series correlation of data features is improved, and the accuracy of trend prediction using the model is improved; thereby providing more accurate data analysis results for the management end of the supply chain, so that the management end can effectively improve each node of the supply chain based on the data analysis results.

[0051] like Figure 2 As shown, an embodiment of the present invention provides a process for training a model, which may include the following steps:

[0052] Step S201: Acquire historical node data of multiple nodes included in the supply chain.

[0053] Specifically, historical node data of multiple nodes included in the supply chain are obtained as training data, for example, historical data within a period of time in the past is obtained, for example, various types of historical node data such as order volume, package volume, visit volume, complaint volume, etc. within 400 days; wherein, in the process of training a tree classification model or a time series model, the historical node data can be divided into a training set and a test set. For example, for the historical node data of the past 400 days, the historical node data of the past 370 days are used as a training set, and the historical node data of the past 30 days are used as a test set. The existing data is compared with the predicted output of the model to evaluate the training effect of the model.

[0054] Step S202: extracting a variety of historical numerical features from the historical node data to iteratively train a tree classification model.

[0055] Step S203: Determine the trained tree classification model as the preset tree classification model.

[0056] Step S204: extracting a plurality of time series features from the plurality of historical numerical features to iteratively train a time series model.

[0057] Step S205: determine the trained tree classification model as the preset tree classification model.

[0058] Specifically, the description of step S202-step S205 is to obtain the historical node data of multiple nodes included in the supply chain; extract a variety of historical numerical features from the historical node data to iteratively train the tree classification model, and determine the trained tree classification model as the preset tree classification model; extract a variety of time series features from the multiple historical numerical features to iteratively train the time series model, and determine the trained time series model as the preset time series model.

[0059] It can be understood that the sequential numbering in the accompanying drawings is only an example, and the tree classification model and the time series model can be trained separately without sequential dependency.

[0060] Further preferably, after step S201 and before step S202 or step S204 of training the model, the training data is screened to improve the training effect of the model, thereby improving the accuracy of the trained preset tree classification model and the preset time series model.

[0061] Specifically, in one embodiment of the present invention, different node levels in the supply chain are used for division and screening. For example, the node levels are divided into three types, ABC (for example, high, medium, and low) according to the node value (for example, determined based on the order volume, profit margin, etc.), and then according to the three corresponding division strategies of ABC, the target historical node data that meets each node level is screened out from the historical node data corresponding to all nodes respectively; and the screened target historical node data is further used to perform the steps of extracting features and training models. That is, after the historical node data of the multiple nodes included in the supply chain are obtained, it also includes: obtaining multiple node levels divided for the multiple nodes included in the supply chain; for each of the node levels, the target historical node data that meets the node level is screened out from the historical node data of the multiple nodes; the iterative training of the tree classification model by extracting multiple numerical features from the historical node data includes: iterative training of the tree classification model by extracting multiple historical numerical features from the target historical node data of the multiple node levels; the iterative training of the time series model by extracting multiple time series features from the multiple numerical features extracted from the historical node data includes: iterative training of the time series model by extracting multiple time series features from the multiple historical numerical features.

[0062] Further, preferably, after screening out the target historical node data that meets the node level from the historical node data of the plurality of nodes, it also includes: when taking order increase as the training target of model training, increasing the proportion of historical numerical features indicating order increase extracted from the target historical node data, and increasing the proportion of time series features indicating order increase extracted from the historical numerical features. Specifically, when training a model (a tree classification model and / or a time series model), different training objectives can be set, for example: the training objective is order increase or order loss; when order increase is used as the training objective for model training, the proportion of historical numerical features indicating order increase extracted from the target historical node data can be increased; historical numerical features indicating order increase include, for example, multiple numerical or time series features such as order volume increase, month-on-month order increase, and growth rate, and the model (a tree classification model and / or a time series model) is trained to focus on model training for predicting order increase; similarly, when order loss is used as the training objective for model training, the proportion of historical numerical features indicating order loss extracted from the target historical node data can be increased; historical numerical features indicating order loss include, for example, multiple numerical or time series features such as order interval length, loss number, loss length, and loss proportion, and the model (a tree classification model and / or a time series model) is trained to focus on model training for predicting order loss.

[0063] The model training method adopted by the embodiment of the present invention combined with the data processing step can greatly improve the value of F1, where F1 is an indicator of the comprehensive evaluation model, and is calculated as 2*P*R / (P+R), where P is the accuracy and R is the recall. This improves the accuracy of the order change trend of the prediction node.

[0064] like Figure 3 As shown, the embodiment of the present invention provides a data processing device 300, including: a data acquisition module 301, an execution prediction module 302 and a data analysis module 303; wherein,

[0065] The data acquisition module 301 is used to acquire node data generated by any node in the supply chain within a set time range;

[0066] The execution prediction module 302 is used to extract a numerical feature matrix including multiple numerical features from the node data, input the numerical feature matrix including multiple numerical features into a preset tree classification model, and use the preset tree classification model to predict the first order change trend of the node; extract a time series feature matrix including multiple time series features from each of the numerical features, input the time series feature matrix including multiple time series features into a preset time series model, and use the preset time series model to predict the second order change trend of the node;

[0067] The data analysis module 303 is used to generate a data analysis result for the node in combination with the first order change trend and the second order change trend, and provide the data analysis result to the management terminal corresponding to the node.

[0068] An embodiment of the present invention also provides an electronic device for data processing, comprising: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in any of the above embodiments.

[0069] An embodiment of the present invention further provides a computer-readable medium on which a computer program is stored. When the program is executed by a processor, the method provided in any of the above embodiments is implemented.

[0070] Figure 4 An exemplary system architecture 400 is shown to which the data processing method or data processing device according to the embodiment of the present invention can be applied.

[0071] like Figure 4As shown, system architecture 400 may include terminal devices 401, 402, 403, network 404 and server 405. Network 404 is used to provide a medium for communication links between terminal devices 401, 402, 403 and server 405. Network 404 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0072] Users can use terminal devices 401, 402, 403 to interact with server 405 through network 404 to receive or send messages, etc. Various client applications can be installed on terminal devices 401, 402, 403, such as e-commerce client applications, web browser applications, search applications, instant messaging tools, and email clients.

[0073] The terminal devices 401 , 402 , and 403 may be various electronic devices having a display screen and supporting various client applications, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0074] Server 405 may be a server that provides various services, such as a background management server that provides support for client applications used by users using terminal devices 401, 402, and 403. The background management server may process received data analysis requests for nodes and feed back the data analysis results of the nodes to the terminal devices.

[0075] It should be noted that the data processing method provided in the embodiment of the present invention is generally executed by the server 405 , and accordingly, the data processing device is generally disposed in the server 405 .

[0076] It should be understood that Figure 4 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0077] Reference below Figure 5 , which shows a schematic diagram of the structure of a computer system 500 of a terminal device suitable for implementing an embodiment of the present invention. Figure 5 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0078] like Figure 5As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the system 500 are also stored. The CPU 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0079] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage section 508 as needed.

[0080] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processing unit (CPU) 501, the above-mentioned functions defined in the system of the present invention are executed.

[0081] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, 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 above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0082] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0083] The modules and / or units involved in the embodiments of the present invention may be implemented in software or hardware. The modules and / or units described may also be set in a processor, for example, they may be described as: a processor including a data acquisition module, a prediction execution module, and a data analysis module; wherein the names of these modules do not constitute a limitation on the module itself in certain circumstances, for example, the data acquisition module may also be described as a "module for acquiring node data generated by any node in the supply chain within a set time range".

[0084] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes: obtaining node data generated by any node in the supply chain within a set time range; extracting a numerical feature matrix including multiple numerical features from the node data, inputting the numerical feature matrix including multiple numerical features into a preset tree classification model, and using the preset tree classification model to predict the first order change trend of the node; extracting a time series feature matrix including multiple time series features from each of the numerical features, inputting the time series feature matrix including multiple time series features into a preset time series model, and using the preset time series model to predict the second order change trend of the node; combining the first order change trend and the second order change trend to generate a data analysis result for the node, and providing the data analysis result to the management end corresponding to the node.

[0085] The embodiments of the present invention can extract numerical features and time series features from the node data of the supply chain, and input the features into a trained preset tree classification model and a trained preset time series model respectively, so as to respectively predict the order change trend of the node, and determine the data analysis result of the node in combination with the obtained order change trend and send it to the management end of the node; the embodiments of the present invention extract different types of correlated features from the data, and perform data analysis in combination with the prediction results of different prediction models, thereby greatly improving the accuracy of data analysis of supply chain nodes and improving the user experience of the management end of managing the supply chain.

[0086] 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 may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data processing method, characterized in that: include: Obtain node data generated by any node in the supply chain within a set time range; Extracting a numerical feature matrix including a plurality of numerical features from the node data, inputting the numerical feature matrix including the plurality of numerical features into a preset tree classification model, and using the preset tree classification model to predict a first order change trend of the node; Extracting a time series feature matrix including a plurality of time series features from each of the numerical features, inputting the time series feature matrix including the plurality of time series features into a preset time series model, and using the preset time series model to predict a change trend of the second order of the node; A data analysis result for the node is generated in combination with the first order change trend and the second order change trend, and the data analysis result is provided to a management terminal corresponding to the node.

2. The method according to claim 1, characterized in that: The step of extracting a numerical feature matrix including a plurality of numerical features from the node data comprises: Acquire data corresponding to the first numerical feature corresponding to one or more unit times from the node data; calculate data corresponding to the second numerical feature of one or more extended unit times based on the data corresponding to the first numerical feature; A numerical feature matrix is ​​constructed using the data corresponding to the first numerical feature and the data corresponding to the second numerical feature.

3. The method according to claim 2, characterized in that The step of extracting a time series feature matrix including a plurality of time series features from each of the numerical features comprises: According to the time sequence, multiple unit times and multiple extended unit times of continuous time are constructed; Extracting a time series feature corresponding to a plurality of time units of continuous time from the data corresponding to the first numerical feature; A time series feature corresponding to a plurality of extended unit times that are continuous in time is extracted from the data corresponding to the second numerical feature.

4. The method according to claim 1, characterized in that: Further including: Acquire historical node data of multiple nodes included in the supply chain; Extracting a plurality of historical numerical features from the historical node data to iteratively train a tree classification model, and determining the trained tree classification model as a preset tree classification model; A plurality of time series features are extracted from the plurality of historical numerical features to iteratively train a time series model, and the trained time series model is determined as a preset time series model.

5. The method according to claim 4, characterized in that After acquiring the historical node data of the multiple nodes included in the supply chain, the method further includes: Acquire multiple node levels classified for multiple nodes included in the supply chain; For each of the node levels, target historical node data satisfying the node level is screened out from the historical node data of the plurality of nodes; The iterative training of the tree classification model by extracting a plurality of numerical features from the historical node data comprises: extracting a plurality of historical numerical features from target historical node data of a plurality of node levels and iteratively training the tree classification model; The method of extracting a plurality of time series features from a plurality of numerical features extracted from the historical node data and iteratively training the time series model includes: extracting a plurality of time series features from the plurality of historical numerical features and iteratively training the time series model.

6. The method according to claim 1, characterized in that Further including: The generating a data analysis result for the node by combining the first order change trend and the first second order change trend includes: The first order change trend and the second order change trend are used to predict order increase or order loss of the node, and generate a data analysis result indicating order increase or order loss.

7. The method according to claim 5 or 6, characterized in that: After selecting the target historical node data that meets the node level from the historical node data of the plurality of nodes, the method further includes: When order increase is used as the training target of model training, the proportion of historical numerical features indicating order increase extracted from the target historical node data is increased, and the proportion of time series features indicating order increase extracted from the historical numerical features is increased.

8. A data processing device, characterized in that: include: Acquisition module, prediction execution module and data analysis module; among them, The data acquisition module is used to acquire node data generated by any node in the supply chain within a set time range; The execution prediction module is used to extract a numerical feature matrix including multiple numerical features from the node data, input the numerical feature matrix including multiple numerical features into a preset tree classification model, and use the preset tree classification model to predict the first order change trend of the node; extract a time series feature matrix including multiple time series features from each of the numerical features, input the time series feature matrix including multiple time series features into a preset time series model, and use the preset time series model to predict the second order change trend of the node; The data analysis module is used to generate a data analysis result for the node in combination with the first order change trend and the second order change trend, and provide the data analysis result to the management end corresponding to the node.

9. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.