Supply chain management method and system based on multi-source data

Abstract

The invention belongs to the technical field of supply chain management, and particularly provides a supply chain management method based on multi-source data, and the method comprises the steps: obtaining the sales data of a retailer in a preset time, and carrying out the preprocessing; analyzing the preprocessed sales data, marking consumers associated with the key factors, and evaluating the influence weights of the key factors on the purchase behaviors of the marked consumers; acquiring and calculating historical sales data and all key factor data within a preset time to obtain enhanced prediction features; constructing a demand prediction model, inputting the preprocessed sales data, the enhanced prediction features, the marked consumers and the evaluated influence weight into the demand prediction model by using a learning algorithm, predicting the supply demand of each node of the supply chain within a preset time through the demand prediction model, and generating a corresponding supply strategy; the supply strategy of the supply chain is flexibly adjusted by integrating multiple data sources, so that quick response to market change and efficient management of the supply chain are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of supply chain management, and particularly relates to a supply chain management method and system based on multi-source data. Background Art

[0002] Supply chain management is a core link in modern enterprise operations, covering the entire process from raw material procurement to the delivery of the final product to consumers. Effective supply chain management can significantly improve enterprise operation efficiency, reduce costs, and enhance customer satisfaction.

[0003] However, traditional supply chain management methods often expose the following limitations when dealing with complex and changing market environments:

[0004] Demand forecasting relies on limited data sources: Traditional methods mainly rely on historical sales data for demand forecasting, ignoring other important factors that may affect consumer purchasing behavior, such as promotional activities, weather changes, and seasonal alternations. This dependence on a single data source leads to a deviation between the forecasting results and the actual market demand, resulting in inaccurate demand forecasting and an inability to respond promptly to market changes.

[0005] In view of the above problems, the present invention aims to provide a supply chain management method and system based on multi-source data, which can flexibly adjust the supply strategy of the supply chain by integrating multiple data sources to achieve a rapid response to market changes and efficient management of the supply chain. Summary of the Invention

[0006] To overcome the deficiencies of the prior art, the present invention provides a supply chain management method and system based on multi-source data to solve the problems in the prior art.

[0007] One embodiment of the present invention provides a supply chain management method based on multi-source data, including the following steps:

[0008] Obtain the sales data of retailers within a preset time and preprocess the sales data;

[0009] Analyze the preprocessed sales data, mark the consumers associated with key factors, and evaluate the influence weight of the key factors on the purchasing behavior of the marked consumers;

[0010] Obtain and calculate the historical sales data and all key factor data within the preset time to obtain enhanced prediction features;

[0011] Construct a demand forecasting model, and input the preprocessed sales data, enhanced prediction features, marked consumers, and evaluated influence weights into the demand forecasting model using a learning algorithm to obtain a trained demand forecasting model;

[0012] Predict the supply demand of each node in the supply chain within a preset time through a trained demand prediction model;

[0013] Generate a supply strategy for each node in the supply chain within a preset time according to the prediction result of the demand prediction model.

[0014] In one embodiment, in the step of analyzing the preprocessed sales data, marking consumers associated with key factors, and evaluating the influence weight of the key factors on the purchase behavior of the marked consumers, the following steps are specifically included:

[0015] Conduct an association analysis on the preprocessed sales data, and use the association rule learning algorithm to determine consumers related to key factors;

[0016] Mark consumers associated with key factors according to the association analysis result;

[0017] Use an influence weight evaluation model to evaluate the influence weight of key factors on the purchase behavior of the marked consumers. Among them, the influence weight is represented by a regression coefficient, and the calculation formula of the regression coefficient is:

[0018]

[0019] Among them, β i represents the influence weight of the i-th key factor, x ij represents the value of the j-th sample on the i-th key factor, y j represents the purchase behavior index of the j-th sample, x ˉ i and y ˉ respectively represent the mean values of the i-th key factor and the purchase behavior index;

[0020] Set a threshold θ. When the influence weight β i exceeds the threshold θ, it is determined that the purchase behavior of the marked consumers is driven by the i-th key factor.

[0021] In one embodiment, in the step of obtaining and calculating the historical sales data and all key factor data within a preset time to obtain enhanced prediction features, the following steps are specifically included:

[0022] Obtain historical sales data and preprocess the historical sales data, and extract several different key factors from the preprocessed historical sales data;

[0023] Calculate the influence weight of several different key factors on the purchase behavior of consumers, and determine the important factors among the several different key factors according to the calculation result of the influence weight;

[0024] Obtain all key factor data within a preset time, and analyze whether there are important factors among all the key factors within the preset time according to the determined important factors;

[0025] If there are important factors, analyze and calculate the occurrence probability of the important factors within the preset time. When the occurrence probability of the important factors is greater than the preset threshold, generate enhanced prediction features.

[0026] In one embodiment, after the step of predicting the supply demand of each node in the supply chain within a preset time through the trained demand prediction model, the following steps are further included:

[0027] Obtain consumer intention data from the labeled consumers, where the consumer intention data is the feedback data of the labeled consumers;

[0028] Determine the repurchase demand of the labeled consumers within the preset time according to the feedback data;

[0029] Dynamically adjust the supply strategy for each node in the supply chain within the preset time based on the repurchase demand.

[0030] In one embodiment, before the step of constructing a demand prediction model and inputting the preprocessed sales data, enhanced prediction features, labeled consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model, the following steps are also included:

[0031] Extract several correlation features from the preprocessed historical sales data and the preprocessed sales data within the preset time;

[0032] Use the association rule algorithm to perform correlation mining on the extracted several correlation features, generate corresponding association rules, evaluate the generated association rules, and filter out strong association rules according to the preset threshold. The association rule is the purchase pattern or relationship between different products shown in the consumer purchase behavior;

[0033] The step of constructing a demand prediction model and inputting the preprocessed sales data, enhanced prediction features, labeled consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model is specifically as follows:

[0034] Construct a demand prediction model, and input the preprocessed sales data, enhanced prediction features, labeled consumers, evaluated influence weights, and strong association rules into the demand prediction model using a learning algorithm to obtain a trained demand prediction model.

[0035] In one embodiment, in the step of using the association rule algorithm to mine the relevance of several extracted association features, generating corresponding association rules, evaluating the generated association rules, and screening out strong association rules according to a preset threshold, the following specific steps are included:

[0036] Based on the extracted association features, determine the minimum support threshold and generate candidate item sets;

[0037] Calculate the support of each candidate item set in the dataset and compare it with the minimum support threshold;

[0038] Screen out the candidate item sets whose support is greater than or equal to the minimum support threshold to obtain frequent item sets;

[0039] Extract all non-empty subsets from the frequent item sets and generate corresponding association rules;

[0040] Calculate the support and confidence of each association rule, and eliminate the association rules that do not meet the minimum support and minimum confidence thresholds to obtain effective association rules;

[0041] Set the minimum thresholds of support, confidence, and lift, and screen out strong association rules.

[0042] In one embodiment, after the step of generating supply strategies for each node of the supply chain within a preset time according to the prediction result of the demand prediction model, the following steps are further included:

[0043] Receive feedback information from each node of the supply chain in real time, and judge whether the predicted supply strategy meets the supply demands of each node of the supply chain according to the received feedback information;

[0044] If the demand is not met, generate a feedback training set according to the received feedback information, and optimize the demand prediction model with the feedback training set.

[0045] This application also relates to a supply chain management system based on multi-source data, including:

[0046] A first data acquisition module for acquiring the sales data of retailers within a preset time;

[0047] A data preprocessing module for preprocessing the sales data;

[0048] A data analysis and evaluation module for analyzing the preprocessed sales data, marking consumers associated with key factors, and evaluating the influence weight of the key factors on the purchase behavior of the marked consumers;

[0049] The second data acquisition module is used to acquire and calculate historical sales data and all key factor data within a preset time period to obtain enhanced prediction features;

[0050] The construction module is used to construct a demand prediction model, and input the preprocessed sales data, enhanced prediction features, labeled consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model;

[0051] The prediction module is used to predict the supply demands of each node in the supply chain within a preset time period through the trained demand prediction model;

[0052] The strategy generation module is used to generate supply strategies for each node in the supply chain within a preset time period according to the prediction results of the demand prediction model.

[0053] This application also relates to a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned supply chain management method based on multi-source data are implemented.

[0054] This application also relates to a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned supply chain management method based on multi-source data are implemented.

[0055] The above-mentioned supply chain management method and system provided by the embodiments have the following beneficial effects:

[0056] 1. By integrating multiple data sources, it predicts how the supply strategy for the supply chain will be in a future period of time, and thus flexibly adjusts the supply strategy of the supply chain to achieve a rapid response to market changes and efficient management of the supply chain.

[0057] 2. In one of the embodiments, by collecting and analyzing the feedback data of labeled consumers, the repeat purchase demand information is extracted, and the supply strategies of each node in the supply chain are dynamically adjusted according to the repeat purchase demand, improving the efficiency and accuracy of supply chain management.

[0058] 3. In one of the embodiments, by judging whether a consumer will purchase another product when buying a product, if so, according to this judgment rule, the supply demand of the supply chain is dynamically adjusted, that is, the original supply chain only supplies one product, and now according to the consumer's purchase habits, the supply demand of the supply chain of another product is also adjusted. Brief Description of the Drawings

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the structures shown in these drawings.

[0060] Figure 1 It is a flowchart of a supply chain management method based on multi-source data provided by an embodiment of the present invention;

[0061] Figure 2 It is a principle block diagram of a computer device provided by an embodiment of the present invention. Specific embodiments

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0063] It should be noted that if there are directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention, the directional indications are only used to explain the relative position relationship and movement conditions between components in a specific posture. If the specific posture changes, the directional indications will also change accordingly.

[0064] In addition, if there are descriptions such as "first" and "second" in the embodiments of the present invention, the descriptions of "first" and "second" are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In addition, if "and / or" or "and / or" appears throughout the text, its meaning includes three parallel solutions. Taking "A and / or B" as an example, it includes solution A, or solution B, or the solution where A and B are satisfied simultaneously. In addition, the technical solutions between the embodiments can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0065] Referring to Figure 1 One embodiment of the present invention provides a supply chain management method based on multi-source data, including the following steps:

[0066] S00. Determine the key factors influencing consumers' purchasing behavior.

[0067] S10. Obtain the sales data of retailers within a preset time and preprocess the sales data.

[0068] S20. Analyze the preprocessed sales data, mark the consumers associated with the key factors, and evaluate the influence weight of the key factors on the purchasing behavior of the marked consumers. Determine whether the purchasing behavior of the marked consumers is driven by the key factors according to the evaluation results. Specifically, it includes the following steps:

[0069] S21. Conduct an association analysis on the preprocessed sales data and use the association rule learning algorithm to determine the consumers related to the key factors;

[0070] S22. Mark the consumers associated with the key factors according to the association analysis results;

[0071] S23. Use an influence weight evaluation model to evaluate the influence weight of the key factors on the purchasing behavior of the marked consumers. Among them, the influence weight is represented by the regression coefficient, and the calculation formula of the regression coefficient is:

[0072]

[0073] where, β i represents the influence weight of the i-th key factor, x ij represents the value of the j-th sample on the i-th key factor, y j represents the purchasing behavior index of the j-th sample, and x̄ i and ȳ respectively represent the means of the i-th key factor and the purchasing behavior index;

[0074] S24. Set a threshold θ. When the influence weight β i exceeds the threshold θ, determine that the purchasing behavior of the marked consumers is driven by the i-th key factor.

[0075] S30. Obtain and calculate the historical sales data and all key factor data within a preset time to obtain enhanced prediction features. Specifically, it includes the following steps:

[0076] S31. Obtain the historical sales data and preprocess the historical sales data, and extract several different key factors from the preprocessed historical sales data;

[0077] S32. Calculate the influence weights of several different key factors on consumers' purchasing behavior, and determine the important factors among the several different key factors according to the calculation results of the influence weights;

[0078] S33. Obtain all key factor data within a preset time, and analyze whether there are important factors among all the key factors within the preset time according to the determined important factors.

[0079] S34. If there are important factors, analyze and calculate the occurrence probability of the important factors within the preset time. When the occurrence probability of the important factors is greater than the preset threshold, generate enhanced prediction features.

[0080] S40. Construct a demand prediction model, and input the preprocessed sales data, enhanced prediction features, labeled consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model.

[0081] S50. Predict the supply demand of each node in the supply chain within a preset time through the trained demand prediction model.

[0082] S60. Generate supply strategies for each node in the supply chain within a preset time according to the prediction results of the demand prediction model.

[0083] As described in step S00 above, the key factors are factors such as promotional activities, seasonality, weather changes, etc., and factors such as price increases, competing products, brand quality problems, etc.; methods such as market research and literature data analysis are used to determine which factors will affect consumers' purchase behavior. For example, when the supply chain is a clothing supply chain, consumers' purchase behavior is most affected by factors such as seasonality, weather changes, and promotional activities; when the supply chain is an electronics supply chain, consumers' purchase behavior is most affected by promotional activity factors.

[0084] As described in step S10 above, the sales data within a preset time is the sales situation of retailers in the past period (such as the past three months), obtained from the retailer's ERP system, POS system, e-commerce platform, etc., and includes information such as product name, sales quantity, sales date, sales price, etc. By using the sales data within a preset time as the prediction basis, the recent purchase trends of consumers can be determined. If historical sales data is used as the prediction basis, there will be more uncertainties. The preprocessing includes data cleaning (removing outliers), missing value processing (such as interpolation or deletion), and data normalization (such as standardization or normalization). The collected sales data is subjected to data cleaning, missing value processing, and data normalization to ensure data quality and make the data suitable for the input requirements of the learning model. By preprocessing the sales data, the sales data can be optimized, which helps to improve the performance and accuracy of the model.

[0085] As described in step S20 above, the pre-processed sales data is analyzed by using an association rule learning algorithm to determine which consumers are affected by the identified key factors, and relevant consumers are labeled according to the analysis results. For example, when the key factor is a promotional activity, we find that the purchase volume of most consumers increases during the promotion; when the key factor is a brand quality problem, we find that the purchase behavior of most consumers decreases due to the brand quality problem of the product. Then these consumers are labeled; then, according to the key factors, an impact weight evaluation model is used to evaluate the impact of the key factors on consumers. By using the labeling method to label the consumers in the sales data within a preset time whose purchase behavior is affected by the key factors, and evaluating the impact weight of the key factors on the purchase behavior of the labeled consumers, it can be determined whether the purchase behavior of the labeled consumers is driven by the key factors according to the evaluation results, which can increase the confidence that the purchase behavior of consumers is affected by the key factors.

[0086] As described in step S30 above, the pre-processing of historical sales data is the same as the pre-processing method of the sales data within a preset time in step S10. By obtaining historical sales data, the impact weights of several different key factors on consumers' purchase behavior in the historical sales data are calculated. According to the impact weights, the impact factor with the largest impact weight is determined as the important factor. For example, in the historical sales data, there are multiple key factors. Calculate which key factor has the greatest impact on consumers' purchase behavior in these sales data, and determine the key factor with the greatest impact as the important factor. And obtain all key factor data within a preset time (including a period of time in the past and a period of time in the future) (all key factor data that will appear in the future period of time, such as promotional activities, season changes, temperature drops, temperature rises, rainfall, etc., and obtain whether there are quality, price increases, competing products, etc. during a period of time). Analyze whether there is an important factor calculated according to the historical sales data among all the obtained key factors. If there is, further analyze the probability that the important factor will occur within the preset time. When the probability of occurrence is greater than the preset threshold, an enhanced prediction feature is obtained according to the important factor and the probability of occurrence of the important factor to increase the accuracy of subsequent prediction results.

[0087] As described in the above steps S40 - S60, the learning algorithm is any one of random forest, support vector machine, neural network or gradient boosting machine. By using the foregoing algorithm, the demand prediction model can learn from the input data and predict the supply demand situation of each node in the supply chain within a preset time. The preset time is a period of time in the future, such as the next one month, two months or three months. For each node in the supply chain: A complete supply chain generally consists of suppliers, manufacturers, distributors, retailers and end - consumers; and each node in the supply chain refers to suppliers, manufacturers, distributors and retailers. The supply demand for each node in the supply chain is determined according to the purchase demand of the end - consumers. When the purchase demand of the end - consumers increases, the supply demand for each node in the supply chain needs to be increased to ensure that each node in the supply chain can keep up with the purchase demand of the end - consumers, so as to respond to market changes in a timely manner. By using the demand prediction model to predict the supply demand situation of each node in the supply chain within a period of time in the future, based on the prediction results of the demand prediction model, corresponding supply strategies for each node in the supply chain within the preset time are generated, and the supply strategies are used to respond to market changes in a timely manner.

[0088] In this embodiment, based on the sales data within the preset time (a past period of time), the influence of each key factor in the historical sales data, all the key factors that occurred within the preset time (including a past period of time and a future period of time) and their occurrence probabilities, the growth or decline trend of consumers' purchase demand in the future period of time is predicted, and corresponding supply strategies are generated accordingly. If there is a growth trend in consumers' purchase demand, the supply demand for each node in the supply chain is increased; if there is a decline trend in consumers' purchase demand, the supply demand for each node in the supply chain is decreased. By integrating multiple data sources, how the supply strategy for the supply chain will be in the future period of time is predicted, and the supply strategy of the supply chain is flexibly adjusted to achieve a rapid response to market changes and efficient management of the supply chain. In this embodiment, the time - series analysis method is not directly used as a node to adjust the supply strategy of the supply chain. There are some drawbacks in using time - series analysis for adjustment: For example, taking the clothing supply chain management as an example, for the north and south regions, the seasons change gradually in the northern region, so the clothing supply management can be carried out according to the time line; while the temperature change in the southern region is unpredictable. If the supply management is carried out according to the time line, it may not be able to respond to the changes in market demand in a timely manner. This embodiment uses a prediction method, which can respond to the changes in market demand in a timely manner according to the actual situation and the needs of end - consumers, and is flexibly adjusted according to different data sources.

[0089] In one of the embodiments, after step S50, the following steps are further included:

[0090] S501. Obtain consumer intention data from the marked consumers, where the consumer intention data is the feedback data of the marked consumers;

[0091] S502. Determine the repurchase demand of the marked consumers within a preset time according to the feedback data;

[0092] S503. Dynamically adjust the supply strategies for each node of the supply chain within the preset time based on the repurchase demand.

[0093] In this embodiment, for the supply strategies of each node of the supply chain, the repurchase demand of the marked consumers is added to further dynamically adjust the supply strategies. For steps S501 and S502, collect the feedback data of the marked consumers through channels such as questionnaires, online evaluations, and customer service records. Preprocess the feedback data, including data cleaning, outlier removal, missing value filling, etc. Use natural language processing (NLP) technology to analyze the text data and extract information such as keywords and sentiment tendencies. According to the purchase history, evaluation content, repurchase intention expression, etc. in the feedback data, identify consumers with a tendency to repurchase. Statistically analyze information such as the repurchased products, repurchase frequencies, and repurchase quantities of these consumers to form a repurchase demand information table. For step S503, compare the extracted repurchase demand information with the demand predicted by the demand prediction model for each node of the supply chain, analyze the differences between the two, identify the deviations between the predicted demand and the actual repurchase demand, and analyze the reasons for the deviations, which may include market changes, changes in consumer behavior, and the impact of promotional activities, etc. According to the analysis results, use the exponential smoothing method to calculate the adjustment coefficient. The specific calculation method is as follows:

[0094] Set the smoothing coefficient α, with a value range between 0 and 1, which is determined according to historical data or experience.

[0095] Calculate the adjustment coefficient ΔD: ΔD = α × (actual repurchase demand - predicted demand) + (1 - α)

[0096] × previous period adjustment coefficient.

[0097] In the initial period, the previous period adjustment coefficient can be set to 0 or estimated according to historical data.

[0098] Apply the adjustment coefficient ΔD to correct the supply plan for each node of the supply chain;

[0099] According to the calculated adjustment coefficient ΔD, correct the supply plan for each node of the supply chain; the correction formula is: corrected supply quantity = original supply quantity + ΔD.

[0100] Ensure that the corrected supply quantity meets the actual repurchase demand, and at the same time avoid overstocking or out-of-stock situations.

[0101] Furthermore, continuously monitor the changes in repurchase demand and readjust the supply strategy as needed.

[0102] Specifically:

[0103] Set up a real-time monitoring system to track the impact of factors such as consumer feedback data and market changes on repurchase demand.

[0104] Periodically re-execute the above steps to update the repurchase demand information and adjustment coefficients based on the latest data.

[0105] Dynamically adjust the supply strategies of each node in the supply chain according to the real-time monitoring results and the updated adjustment coefficients.

[0106] Ensure that the supply strategy is always consistent with the actual repurchase demand, and improve the response speed and flexibility of the supply chain.

[0107] In one embodiment, before step S40, the following steps are further included:

[0108] S401. Extract several correlation features from the preprocessed historical sales data and the sales data within a preset time period after preprocessing;

[0109] S402. Use the association rule algorithm to perform correlation mining on the extracted several correlation features, generate corresponding association rules, evaluate the generated association rules, and filter out strong association rules according to a preset threshold. The association rules are the purchase patterns or relationships between different products shown in consumers' purchase behaviors. Specifically, the following steps are included:

[0110] S4021. Based on the extracted correlation features, determine the minimum support threshold and generate candidate item sets;

[0111] S4022. Calculate the support of each candidate item set in the dataset and compare it with the minimum support threshold;

[0112] S4023. Filter out the candidate item sets whose support is greater than or equal to the minimum support threshold to obtain frequent item sets;

[0113] S4024. Extract all non-empty subsets from the frequent item sets and generate corresponding association rules;

[0114] S4025. Calculate the support and confidence of each association rule, and eliminate the association rules that do not meet the minimum support and minimum confidence thresholds to obtain effective association rules;

[0115] S4026. Set the minimum thresholds for support, confidence, and lift, and filter out strong association rules.

[0116] Based on the above solution, step S40 is specifically:

[0117] Construct a demand forecasting model. Input the pre - processed sales data, enhanced forecasting features, labeled consumers, evaluated impact weights, and strong association rules into the demand forecasting model using a learning algorithm to obtain a trained demand forecasting model.

[0118] In this embodiment, by determining whether a consumer will purchase another product when buying one product, if so, according to this judgment rule, dynamically adjust the supply demand of the supply chain. That is, the original supply chain only supplies one product, and now according to the consumer's purchase habits, also adjust the supply demand of the supply chain of another product. For example, the supply chain is an electronic product supply chain. We find that most consumers will buy corresponding accessories for protection, use, etc. when buying a mobile phone / computer. Here, in a future period, while adjusting the original supply chain according to the consumer's purchase demand situation, also dynamically adjust the corresponding supply chain to enhance the consumer's purchase experience.

[0119] In one of the embodiments, after step S60, the following steps are further included:

[0120] S70. Real - time receive feedback information from each node of the supply chain, and judge whether the predicted supply strategy meets the supply demand of each node of the supply chain according to the received feedback information;

[0121] S80. If the demand is not met, generate a feedback training set according to the received feedback information, and optimize the demand forecasting model with the feedback training set.

[0122] In this embodiment, the feedback information can be obtained by predicting the supply strategies of each node of the supply chain and generating corresponding supply strategies, then sending out questionnaires to each node of the supply chain, and obtaining the feedback information according to the content of the questionnaires feedback by each node. According to the content of the feedback information, judge whether the supply strategy generated by this prediction is accurate; if there is dissatisfied feedback content, generate a feedback training set according to the feedback content, and then input the feedback training set into the demand forecasting model for optimization training, so as to continuously correct the prediction scheme for the supply strategy and improve the accuracy and efficiency of supply chain management.

[0123] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0124] In one embodiment, a supply chain management system based on multi-source data is provided, which corresponds to a supply chain management method based on multi-source data in the above embodiment. The supply chain management system based on multi-source data includes:

[0125] A first data acquisition module for acquiring the sales data of retailers within a preset time;

[0126] A data preprocessing module for preprocessing the sales data;

[0127] A data analysis and evaluation module for analyzing the preprocessed sales data, marking consumers associated with key factors, and evaluating the influence weight of the key factors on the purchasing behavior of the marked consumers;

[0128] A second data acquisition module for acquiring and calculating the historical sales data and all key factor data within a preset time to obtain enhanced prediction features;

[0129] A construction module for constructing a demand prediction model, and inputting the preprocessed sales data, enhanced prediction features, marked consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model;

[0130] A prediction module for predicting the supply demand of each node in the supply chain within a preset time through the trained demand prediction model;

[0131] A strategy generation module for generating supply strategies for each node in the supply chain within a preset time according to the prediction results of the demand prediction model.

[0132] Optionally, it further includes:

[0133] An adjustment module for dynamically adjusting the supply strategies for each node in the supply chain within a preset time based on the repeat purchase demand.

[0134] Optionally, it further includes:

[0135] An associated feature extraction module for extracting several associated features from the preprocessed historical sales data and the preprocessed sales data within a preset time.

[0136] Optionally, it further includes:

[0137] An information receiving module for real-time receiving feedback information from each node in the supply chain, and judging whether the predicted supply strategy meets the supply demand of each node in the supply chain according to the received feedback information.

[0138] For the specific limitations of a supply chain management system based on multi-source data, reference can be made to the limitations of a supply chain management method based on multi-source data in the above text, which will not be elaborated here. Each module in the above supply chain management system based on multi-source data can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0139] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 2 shown. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store sales data, data processing, predictive analysis, etc. The network interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a supply chain management method based on multi-source data.

[0140] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it implements a supply chain management method based on multi-source data.

[0141] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, it implements a supply chain management method based on multi-source data.

[0142] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0143] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0144] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A supply chain management method based on multi-source data, characterized in that, It includes the following steps: Obtain the sales data of the retailer within a preset time, and preprocess the sales data; Analyze the preprocessed sales data, mark the consumers associated with the key factors, and evaluate the influence weight of the key factors on the purchase behavior of the marked consumers; Obtain and calculate the historical sales data and all key factor data within the preset time to obtain enhanced prediction features; Construct a demand forecasting model, and input the preprocessed sales data, enhanced prediction features, marked consumers, and evaluated influence weights into the demand forecasting model using a learning algorithm to obtain a trained demand forecasting model; Predict the supply demand of each node in the supply chain within the preset time through the trained demand forecasting model; Generate a supply strategy for each node in the supply chain within the preset time according to the prediction results of the demand forecasting model.

2. The supply chain management method based on multi-source data according to claim 1, wherein, In the step of analyzing the preprocessed sales data, marking the consumers associated with the key factors, and evaluating the influence weight of the key factors on the purchase behavior of the marked consumers, it specifically includes the following steps: Conduct an association analysis on the preprocessed sales data, and use the association rule learning algorithm to determine the consumers related to the key factors; Mark the consumers associated with the key factors according to the association analysis results; Adopt an influence weight evaluation model to evaluate the influence weight of the key factors on the purchase behavior of the marked consumers. Among them, the influence weight is represented by a regression coefficient, and the calculation formula of the regression coefficient is: Among them, β i represents the influence weight of the i-th key factor, x ij represents the value of the j-th sample on the i-th key factor, y j represents the purchase behavior index of the j-th sample, x ˉ i and y ˉ respectively represent the means of the i-th key factor and the purchase behavior index; Set a threshold θ. When the influence weight β i exceeds the threshold θ, it is determined that the purchase behavior of the marked consumers is driven by the i-th key factor.

3. The supply chain management method based on multi-source data according to claim 2, characterized in that, In the step of obtaining and calculating the historical sales data and all key factor data within the preset time to obtain enhanced prediction features, it specifically includes the following steps: Obtain the historical sales data and preprocess the historical sales data, and extract several different key factors from the preprocessed historical sales data; Calculate the influence weight of several different key factors on the purchase behavior of consumers, and determine the important factors among the several different key factors according to the calculation results of the influence weight; Obtain all key factor data within the preset time, and analyze whether there are important factors among all key factors within the preset time according to the determined important factors; If there are important factors, analyze and calculate the occurrence probability of the important factors within the preset time. When the occurrence probability of the important factors is greater than the preset threshold, generate enhanced prediction features.

4. The supply chain management method based on multi-source data according to claim 1, characterized in that After the step of predicting the supply demand of each node in the supply chain within the preset time through the trained demand forecasting model, it further includes the following steps: Obtain consumer intention data from the marked consumers, and the consumer intention data is the feedback data of the marked consumers; Determine the repurchase demand of the marked consumers within the preset time according to the feedback data; Dynamically adjust the supply strategy for each node in the supply chain within the preset time based on the repurchase demand.

5. The supply chain management method based on multi-source data according to claim 3, characterized in that Before the step of constructing a demand forecasting model, inputting the preprocessed sales data, enhanced prediction features, marked consumers, and evaluated influence weights into the demand forecasting model using a learning algorithm to obtain a trained demand forecasting model, it also includes the step: Extract several associated features from the preprocessed historical sales data and the preprocessed sales data within a preset time period; Use the association rule algorithm to perform association mining on the several extracted associated features, generate corresponding association rules, evaluate the generated association rules, and filter out strong association rules according to a preset threshold. The association rule is the purchase pattern or relationship between different products shown in consumers' purchase behaviors; The steps of constructing a demand prediction model by inputting the preprocessed sales data, enhanced prediction features, labeled consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model are specifically as follows: Construct a demand prediction model, and input the preprocessed sales data, enhanced prediction features, labeled consumers, evaluated influence weights, and strong association rules into the demand prediction model using a learning algorithm to obtain a trained demand prediction model.

6. The supply chain management method based on multi-source data according to claim 5, characterized in that In the step of using the association rule algorithm to perform association mining on the several extracted associated features, generate corresponding association rules, evaluate the generated association rules, and filter out strong association rules according to a preset threshold, it specifically includes the following steps: Based on the extracted associated features, determine the minimum support threshold and generate candidate item sets; Calculate the support of each candidate item set in the dataset and compare it with the minimum support threshold; Filter out the candidate item sets whose support is greater than or equal to the minimum support threshold to obtain frequent item sets; Extract all non-empty subsets from the frequent item sets and generate corresponding association rules; Calculate the support and confidence of each association rule, and eliminate the association rules that do not meet the minimum support and minimum confidence thresholds to obtain valid association rules; Set the minimum thresholds for support, confidence, and lift, and filter out strong association rules.

7. The supply chain management method based on multi-source data according to claim 1, characterized in that, After the step of generating supply strategies for each node of the supply chain within a preset time according to the prediction result of the demand prediction model, the following steps are further included: Receive feedback information from each node of the supply chain in real time, and judge whether the predicted supply strategy meets the supply demands of each node of the supply chain according to the received feedback information; If the demand is not met, generate a feedback training set according to the received feedback information, and optimize the demand prediction model with the feedback training set.

8. A supply chain management system based on multi-source data, which is used to implement the steps of a supply chain management method based on multi-source data as described in any one of claims 1-7, characterized in that, Include: A first data acquisition module for acquiring the sales data of a retailer within a preset time period; A data preprocessing module for preprocessing the sales data; A data analysis and evaluation module for analyzing the preprocessed sales data, marking the consumers associated with key factors, and evaluating the influence weights of the key factors on the purchase behaviors of the marked consumers; A second data acquisition module for acquiring and calculating the historical sales data and all key factor data within a preset time period to obtain enhanced prediction features; A construction module for constructing a demand prediction model, and inputting the preprocessed sales data, enhanced prediction features, labeled consumers, and evaluated influence weights into the demand prediction model using a learning algorithm to obtain a trained demand prediction model; A prediction module for predicting the supply demand of each node in the supply chain within a preset time through a trained demand prediction model; A strategy generation module for generating supply strategies for each node in the supply chain within a preset time according to the prediction results of the demand prediction model.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of a supply chain management method based on multi-source data according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of a supply chain management method based on multi-source data according to any one of claims 1-7.

Images