A sales big data analysis method and system based on cross-border e-commerce ERP

Abstract

The present invention relates to the field of data analysis technology, and invents a sales big data analysis method based on cross-border e-commerce ERP, including: extracting enterprise resource data from an enterprise management platform, and extracting an inventory resource feature set from the enterprise resource data; performing clustering dimensionality reduction and correlation dimensionality reduction on the inventory resource feature set to obtain a product resource feature set; performing autocorrelation analysis and time series feature extraction on the product resource features to obtain product time series features; forward propagating the product time series features to obtain analysis product data, and extracting unsalable product inventory from the analysis product data; extracting product sales features corresponding to the unsalable product inventory from the enterprise resource data, and performing sales strategy analysis on the unsalable product inventory according to the product sales features to obtain product sales strategies. The present invention also proposes a sales big data analysis system based on cross-border e-commerce ERP. The present invention can improve the accuracy of sales analysis.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a sales big data analysis method and system based on cross-border e-commerce ERP. Background Art

[0002] Enterprise Resource Planning (ERP) is a management software system designed specifically for enterprises to help them integrate and optimize all aspects of their business. ERP can integrate and automate core businesses such as supply chain management, logistics, finance, inventory, order processing, customer management, product management, and procurement in enterprise operations, effectively improve business efficiency, reduce operating costs, and enhance the company's ability to respond to the market.

[0003] In order to optimize inventory management and improve enterprise operational efficiency, it is necessary to conduct sales analysis on the enterprise. Existing sales analysis methods usually rely on internal sales records of the enterprise. In actual applications, it is difficult to combine with external time data, resulting in a single analysis perspective and an inability to fully reflect market changes. There is a certain lag, which may lead to low accuracy in sales analysis. Summary of the invention

[0004] The present invention provides a sales big data analysis method and system based on cross-border e-commerce ERP, the main purpose of which is to solve the problem of low accuracy in sales analysis.

[0005] To achieve the above purpose, the present invention provides a sales big data analysis method based on cross-border e-commerce ERP, comprising:

[0006] Extracting data from the enterprise management platform to obtain enterprise resource data, and extracting inventory resource feature sets from the enterprise resource data;

[0007] The inventory resource feature set is clustered and reduced in dimension and correlation to obtain a product resource feature set, wherein the inventory resource feature set is clustered and reduced in dimension and correlation to obtain a product resource feature set, including: performing central feature screening on the inventory resource feature set according to Gaussian distribution to obtain a primary central resource feature group; performing neighborhood division on the primary central resource feature group according to the inventory resource feature set to obtain a primary resource feature group set; selecting primary resource feature groups in the primary resource feature group set one by one as target primary resource feature groups, and using the primary central resource features corresponding to the target primary resource feature groups in the primary central resource feature group as target primary central resource features; selecting inventory resource features in the inventory resource feature set one by one as target inventory resource features, and calculating the product feature distance between the target inventory resource features and the target primary resource feature group according to the target primary central resource features using the following product feature distance algorithm: in, refers to the characteristic distance of the product, is the maximum value symbol, is the feature number, is the total number of features of the target primary resource feature group, is the first in the target primary resource feature group The weight of the inventory resource characteristics, is the first in the target primary resource feature group Inventory resource characteristics, is the target inventory resource characteristic, is the target inventory resource characteristic The weight of is the inner product symbol, is the modulo symbol, is the target primary center resource characteristic, is the weight of the target primary central resource feature; generating a product feature distance set according to the product feature distances of all target inventory resource features in the inventory resource feature set, and updating the target primary resource feature group into a target secondary resource feature group according to the product feature distance set; generating a secondary resource feature group set according to all target secondary resource feature groups, iteratively updating the secondary resource feature group set to obtain a product resource feature set; performing correlation dimension reduction on the product resource feature set to obtain a product resource feature set;

[0008] Performing autocorrelation analysis on the product resource feature set to obtain product hysteresis features, and performing time series feature extraction on the product resource feature set according to the product hysteresis features to obtain product time series features;

[0009] Perform forward propagation on the product time series features to obtain analysis product data, and extract slow-moving product inventory from the analysis product data;

[0010] The product sales characteristics corresponding to the slow-moving product inventory are extracted from the enterprise resource data, and a sales strategy analysis is performed on the slow-moving product inventory based on the product sales characteristics to obtain a product sales strategy.

[0011] Optionally, extracting the inventory resource feature set from the enterprise resource data includes:

[0012] Extracting data attributes from the enterprise resource data to obtain a resource data attribute set;

[0013] Performing inventory keyword matching on the resource data attribute set to obtain an inventory resource attribute set;

[0014] extracting an inventory resource data set from the enterprise resource data according to the inventory resource attribute set;

[0015] Performing data standardization and date marking operations on the inventory resource data set to obtain a standard inventory data set;

[0016] Feature extraction and feature weighting are performed on the standard inventory data set to obtain an inventory resource feature set.

[0017] Optionally, performing correlation dimension reduction on the product resource feature set to obtain the product resource feature set includes:

[0018] Performing feature standardization operations on each product resource feature class in the product resource feature class set to obtain a standard product feature class set;

[0019] The product correlation between each standard product feature class in the standard product feature class set is calculated using the following product correlation algorithm: in, Refers to the standard product feature category set of standard product feature categories and standard product feature classes product relevance, , is the serial number of the standard product feature class in the standard product feature class set, It is the standard product feature class The total number of features, It is the standard product feature class The total number of features, , is the feature number, is the covariance operator symbol, is the standard deviation operator, Refers to the standard product feature class The Features, Refers to the standard product feature class The Features

[0020] Filtering out relevant feature group sets from the standard product feature set according to the product relevance;

[0021] Merging related features of the related feature cluster group set to obtain a merged feature cluster set;

[0022] updating the standard product feature set into a decorrelated feature set according to the merged feature set;

[0023] The principal component dimension reduction is performed on each decorrelation feature class in the decorrelation feature class set to obtain a product resource feature set.

[0024] Optionally, performing autocorrelation analysis on the product resource feature set to obtain product hysteresis features includes:

[0025] Sorting the product resource feature set in time sequence to obtain a product feature sequence;

[0026] Performing self-attention encoding on the product feature sequence to obtain a self-attention weight;

[0027] Extracting the lag period of the product feature sequence according to the self-attention weight to obtain a lag coefficient;

[0028] The following autocorrelation algorithm is used to perform autocorrelation analysis on the product feature sequence according to the lag coefficient to obtain the autocorrelation coefficient: in, The hysteresis coefficient is The autocorrelation coefficient when refers to the hysteresis coefficient, is the feature number, is the total number of features of the product feature sequence, is the first in the product feature sequence Features, is the first in the product feature sequence Features

[0029] A product lag characteristic is generated based on the autocorrelation coefficient.

[0030] Optionally, extracting a time series feature from the product resource feature set according to the product hysteresis feature to obtain a product time series feature includes:

[0031] generating a product feature sequence according to the product resource feature set;

[0032] Performing a stationarity check on the product feature sequence to obtain a feature stationarity coefficient;

[0033] Performing a differential operation on the product feature sequence according to the feature stability coefficient to obtain a differential product feature sequence;

[0034] Performing moving average analysis on the differential product feature sequence to obtain partial regression features;

[0035] The time series characteristics of the differential product feature sequence are fitted according to the product hysteresis characteristics and the partial regression characteristics to obtain product time series characteristics.

[0036] Optionally, extracting the slow-moving product inventory from the analyzed product data includes:

[0037] Splitting the analyzed product data into product data sets according to preset product categories;

[0038] Performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set;

[0039] Using a preset slow-selling threshold value to perform slow-selling statistics on the product inventory quantity set to obtain a slow-selling product quantity set;

[0040] Generate a slow-moving product inventory according to the slow-moving product quantity set and the product data set.

[0041] Optionally, performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set includes:

[0042] Selecting product data in the product data set one by one as target product data, and extracting product supply chain data, product initial inventory data and product logistics data from the target product data;

[0043] Extracting product warehousing data from the product supply chain data and the product logistics data;

[0044] Extracting product delivery data from the product logistics data;

[0045] Calculate the product inventory quantity according to the product entry data, the product exit data and the product initial inventory data;

[0046] The product inventory quantities of all target product data in the product data set are aggregated into a product inventory quantity set.

[0047] Optionally, extracting product sales features corresponding to the slow-moving product inventory from the enterprise resource data includes:

[0048] Extracting slow-moving product category characteristics from the slow-moving product inventory;

[0049] Extracting product sales data from the enterprise resource data according to the characteristics of the slow-selling product categories;

[0050] Extracting a product user group from the product sales data, and performing feature extraction on the product user group to obtain product user features;

[0051] Performing sales cycle analysis on the product sales data to obtain sales cycle characteristics;

[0052] Extracting product price features from the product sales data;

[0053] Product sales characteristics are generated according to the product price characteristics, the product user characteristics and the sales cycle characteristics.

[0054] Optionally, performing a sales strategy analysis on the slow-moving product inventory according to the product sales characteristics to obtain a product sales strategy includes:

[0055] Extracting the slow-moving product category characteristics and slow-moving quantity characteristics from the slow-moving product inventory;

[0056] Performing sales feature splitting on the product sales feature, the unsalable product category feature, and the unsalable quantity feature to obtain a sales feature set;

[0057] Performing regression analysis on the sales feature set to obtain standard sales features;

[0058] Perform strategy mapping on the standard sales features to obtain product sales strategies.

[0059] In order to solve the above problems, the present invention also provides a sales big data analysis system based on cross-border e-commerce ERP, the system comprising:

[0060] A feature extraction module is used to extract data from the enterprise management platform, obtain enterprise resource data, and extract inventory resource feature sets from the enterprise resource data;

[0061] The correlation dimension reduction module is used to perform clustering dimension reduction and correlation dimension reduction on the inventory resource feature set to obtain a product resource feature set, wherein the clustering dimension reduction and correlation dimension reduction on the inventory resource feature set to obtain the product resource feature set includes: performing central feature screening on the inventory resource feature set according to Gaussian distribution to obtain a primary central resource feature group; performing neighborhood division on the primary central resource feature group according to the inventory resource feature set to obtain a primary resource feature group set; selecting primary resource feature groups in the primary resource feature group set as target primary resource feature groups one by one, and using the primary central resource features corresponding to the target primary resource feature groups in the primary central resource feature group as target primary central resource features; selecting inventory resource features in the inventory resource feature set one by one as target inventory resource features, and calculating the product feature distance between the target inventory resource features and the target primary resource feature group according to the target primary central resource features using the following product feature distance algorithm: in, refers to the characteristic distance of the product, is the maximum value symbol, is the feature number, is the total number of features of the target primary resource feature group, is the first in the target primary resource feature group The weight of the inventory resource characteristics, is the first in the target primary resource feature group Inventory resource characteristics, is the target inventory resource characteristic, is the target inventory resource characteristic The weight of is the inner product symbol, is the modulo symbol, is the target primary center resource characteristic, is the weight of the target primary central resource feature; generating a product feature distance set according to the product feature distances of all target inventory resource features in the inventory resource feature set, and updating the target primary resource feature group into a target secondary resource feature group according to the product feature distance set; generating a secondary resource feature group set according to all target secondary resource feature groups, iteratively updating the secondary resource feature group set to obtain a product resource feature set; performing correlation dimension reduction on the product resource feature set to obtain a product resource feature set;

[0062] A time series analysis module, used for performing autocorrelation analysis on the product resource feature set to obtain product hysteresis features, and performing time series feature extraction on the product resource feature set according to the product hysteresis features to obtain product time series features;

[0063] A slow-selling analysis module, used for forward propagating the product time series features to obtain analysis product data, and extracting the slow-selling product inventory from the analysis product data;

[0064] The sales analysis module is used to extract the product sales characteristics corresponding to the slow-moving product inventory from the enterprise resource data, and perform sales strategy analysis on the slow-moving product inventory based on the product sales characteristics to obtain a product sales strategy.

[0065] The present invention can perform resource inventory data analysis from multiple dimensions by performing data warehouse extraction and inventory resource feature analysis, thereby improving the comprehensiveness of data during sales analysis and thus improving the accuracy of sales analysis; by performing clustering dimensionality reduction and correlation dimensionality reduction, it can reduce the feature dimensions of inventory resource features, remove redundancy and multicollinearity between features, and improve the efficiency of big data analysis; by performing autocorrelation analysis and time series feature extraction, it can discover the periodicity or trend correlation of product data, facilitate subsequent predictive analysis of product data in future time periods, and improve the accuracy of sales data analysis.

[0066] By performing forward propagation and extracting the inventory of unsalable products, the changes in product inventory in the future time period can be analyzed, thereby improving the accuracy of product sales analysis, and unsalable products can be screened out to facilitate subsequent targeted sales analysis. By extracting product sales characteristics and performing sales strategy analysis, suitable promotion strategies can be analyzed based on the historical sales data of products that may be unsalable for sales analysis, thereby improving the accuracy of sales analysis. Therefore, the sales big data analysis method and system based on cross-border e-commerce ERP proposed in the present invention can solve the problem of low accuracy when performing sales analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A flowchart of a sales big data analysis method based on cross-border e-commerce ERP provided by one embodiment of the present invention;

[0068] Figure 2 A schematic flow chart of a method for performing correlation dimensionality reduction provided in one embodiment of the present invention;

[0069] Figure 3 A schematic flow chart of a method for performing autocorrelation analysis provided by one embodiment of the present invention;

[0070] Figure 4 A functional module diagram of a sales big data analysis system based on cross-border e-commerce ERP provided by one embodiment of the present invention;

[0071] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0072] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0073] The embodiment of the present application provides a sales big data analysis method based on cross-border e-commerce ERP. The execution subject of the sales big data analysis method based on cross-border e-commerce ERP includes but is not limited to at least one of the electronic devices such as the server, the terminal, etc. that can be configured to execute the method provided in the embodiment of the present application. In other words, the sales big data analysis method based on cross-border e-commerce ERP can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0074] Reference Figure 1 FIG. 1 is a flow chart of a sales big data analysis method based on cross-border e-commerce ERP provided by an embodiment of the present invention. In this embodiment, the sales big data analysis method based on cross-border e-commerce ERP includes:

[0075] S1. Extracting the data warehouse of the enterprise management platform to obtain enterprise resource data, and extracting an inventory resource feature set from the enterprise resource data.

[0076] In detail, the enterprise management platform can be the Enterprise Resource Planning (ERP) of cross-border e-commerce, which is a management software system specially designed for enterprises to help enterprises integrate and optimize all aspects of their business. The enterprise resource planning can realize the integration and automation of core businesses such as supply chain management, logistics, finance, inventory, order processing, customer management, product management, procurement, etc. in enterprise operations, effectively improve business efficiency, reduce operating costs, and enhance the company's responsiveness to the market.

[0077] Specifically, the data warehouse extraction refers to the use of the data warehouse to extract relevant data of the enterprise management platform. The data warehouse is a historical data storage system that integrates multiple sources and is mainly used to support business intelligence and data analysis. Unlike the real-time data in the enterprise management platform, the data warehouse focuses on the organization, aggregation and analysis of historical data, and is suitable for management and analysts to conduct long-term trend analysis, forecasting, and strategic decision-making.

[0078] In detail, the enterprise resources include resource data related to product sales, order processing, product inventory, product logistics, financial data and supply chain data of the enterprise management platform.

[0079] In the embodiment of the present invention, extracting the inventory resource feature set from the enterprise resource data includes:

[0080] Extracting data attributes from the enterprise resource data to obtain a resource data attribute set;

[0081] Performing inventory keyword matching on the resource data attribute set to obtain an inventory resource attribute set;

[0082] extracting an inventory resource data set from the enterprise resource data according to the inventory resource attribute set;

[0083] Performing data standardization and date marking operations on the inventory resource data set to obtain a standard inventory data set;

[0084] Feature extraction and feature weighting are performed on the standard inventory data set to obtain an inventory resource feature set.

[0085] In detail, the data attribute extraction refers to extracting the attributes corresponding to each data in the enterprise resource data, that is, the corresponding module function attributes of each data in the enterprise management platform. The various resource data attributes in the resource data attribute set include supply chain attributes, logistics attributes, financial attributes, inventory attributes, order processing attributes, and customer management attributes, etc.

[0086] Specifically, the inventory keyword matching refers to matching the resource data attributes related to the inventory keyword in the resource data attribute set as resource inventory attributes, and the inventory keyword may be "inventory", "warehouse in", and "warehouse out".

[0087] In detail, the data standardization refers to operations such as data cleaning, default data filling, and unified data format of the inventory resource data. Data standardization can be performed using Python's Pandas library, boxplot method, Z-score standardization, or Min-Max standardization.

[0088] Specifically, the feature extraction refers to performing data vectorization or data feature calculation operations on the standard inventory data, such as performing vectorization operations on the inventory quantity, warehousing time, warehousing time and supply chain cycle time of each product, and calculating the inventory turnover rate and replenishment frequency and other characteristics of each product. The feature weighting refers to weighting according to the inventory resource attributes of each standard inventory data according to preset weight rules or statistical frequency and other methods.

[0089] In the embodiment of the present invention, by performing data warehouse extraction and inventory resource characteristics, resource inventory data analysis can be performed from multiple dimensions, thereby improving the comprehensiveness of data during sales analysis and thus improving the accuracy of sales analysis.

[0090] S2. Perform clustering dimensionality reduction and correlation dimensionality reduction on the inventory resource feature set to obtain a product resource feature set.

[0091] In the embodiment of the present invention, the step of performing clustering dimensionality reduction and correlation dimensionality reduction on the inventory resource feature set to obtain a product resource feature set includes:

[0092] Performing central feature screening on the inventory resource feature set according to Gaussian distribution to obtain a primary central resource feature group;

[0093] Performing neighborhood division on the primary central resource feature group according to the inventory resource feature set to obtain a primary resource feature group set;

[0094] Selecting primary resource feature groups in the primary resource feature group set one by one as target primary resource feature groups, and taking the primary central resource feature corresponding to the target primary resource feature group in the primary central resource feature group as the target primary central resource feature;

[0095] Selecting inventory resource features in the inventory resource feature set one by one as target inventory resource features, and calculating the product feature distance between the target inventory resource feature and the target primary resource feature group according to the target primary central resource feature;

[0096] Generate a product feature distance set according to the product feature distances of all target inventory resource features in the inventory resource feature set, and update the target primary resource feature group into a target secondary resource feature group according to the product feature distance set;

[0097] Generating a secondary resource feature group set according to all target secondary resource feature groups, iteratively updating the secondary resource feature group set to obtain a product resource feature class set;

[0098] The product resource feature set is subjected to correlation dimension reduction to obtain a product resource feature set.

[0099] In detail, the central feature screening of the inventory resource feature set according to the Gaussian distribution includes calculating the time-series Gaussian distribution of the inventory resource feature set, and screening out several inventory resource features with a long time-series Gaussian distribution distance from the inventory resource feature set as primary central resource features and aggregating them into a primary central resource feature group according to the time-series Gaussian distribution.

[0100] Specifically, the neighborhood division refers to allocating each inventory resource feature in the inventory resource feature set to the feature group where each primary central resource feature in the primary central resource feature group is located according to the neighborhood proximity principle to obtain a primary resource feature group set.

[0101] In detail, the product feature distance is calculated using the following product feature distance algorithm: in, refers to the characteristic distance of the product, is the maximum value symbol, is the feature number, is the total number of features of the target primary resource feature group, is the first in the target primary resource feature group The weight of the inventory resource characteristics, is the first in the target primary resource feature group Inventory resource characteristics, is the target inventory resource characteristic, is the target inventory resource characteristic The weight of is the inner product symbol, is the modulo symbol, is the target primary center resource characteristic, is the weight of the target primary center resource feature.

[0102] Specifically, the product feature distance algorithm can calculate feature distances based on the weighted feature distances between the target inventory resource features and the target primary resource feature group and the target primary central resource features and the cosine feature distance algorithm between features, thereby improving the accuracy of clustering.

[0103] In detail, updating the target primary resource feature group into a target secondary resource feature group according to the product feature distance set means allocating each feature in the target primary resource feature group to the feature group where the nearest primary central resource feature is located according to the proximity principle and the product feature distances in the product feature distance set, and the KNN clustering algorithm can be used for iterative updating.

[0104] Specifically, refer to Figure 2 As shown, the correlation dimension reduction is performed on the product resource feature set to obtain a product resource feature set, including:

[0105] S21, performing feature standardization operation on each product resource feature class in the product resource feature class set to obtain a standard product feature class set;

[0106] S22, calculating the product correlation between each standard product feature class in the standard product feature class set;

[0107] S23, selecting a relevant feature group set from the standard product feature set according to the product relevance;

[0108] S24, merging relevant features of the relevant feature cluster set to obtain a merged feature cluster set;

[0109] S25, updating the standard product feature set into a decorrelated feature set according to the merged feature set;

[0110] S26. Perform principal component dimensionality reduction on each decorrelation feature class in the decorrelation feature class set to obtain a product resource feature set.

[0111] In detail, the feature standardization refers to updating the features in each product resource feature class to a form with a mean value of 0 and a standard deviation of 1.

[0112] Specifically, screening out the relevant feature class set means aggregating the standard product feature classes whose product correlation in the standard product feature class set is greater than a preset correlation threshold into a relevant feature class group, and aggregating all relevant feature class groups into a relevant feature class set.

[0113] In detail, the product relevance is calculated using the following product relevance algorithm: in, Refers to the standard product feature category set of standard product feature categories and standard product feature classes product relevance, , is the serial number of the standard product feature class in the standard product feature class set, It is the standard product feature class The total number of features, It is the standard product feature class The total number of features, , is the feature number, is the covariance operator symbol, is the standard deviation operator, Refers to the standard product feature class The Features, Refers to the standard product feature class The Features.

[0114] Specifically, the product correlation algorithm can determine the correlation between two feature classes based on the average inter-class correlation of the features, thereby improving the representation of each product resource feature after feature dimensionality reduction, and can merge relevant features by screening out redundant classes or linear combination.

[0115] Specifically, principal component analysis or linear merging methods may be used to perform principal component dimensionality reduction on each decorrelation feature class in the decorrelation feature class set to obtain a product resource feature set.

[0116] In detail, the feature screening refers to screening out several features with larger decorrelation feature values ​​in the decorrelation feature vector, the feature splicing refers to splicing the screened features into a product feature matrix, and the feature mapping refers to multiplying the product feature matrix by the target decorrelation feature matrix.

[0117] In the embodiment of the present invention, by performing clustering dimensionality reduction and correlation dimensionality reduction, the feature dimensions of inventory resource features can be reduced, redundancy and multicollinearity between features can be removed, and the efficiency of big data analysis can be improved.

[0118] S3. Perform autocorrelation analysis on the product resource feature set to obtain product hysteresis features, and perform time series feature extraction on the product resource feature set based on the product hysteresis features to obtain product time series features.

[0119] In detail, the product hysteresis feature refers to the feature of the mutual influence between various features of the product resource feature, and the product hysteresis feature can improve the accuracy of time series prediction analysis.

[0120] In the embodiment of the present invention, refer to Figure 3 As shown, the autocorrelation analysis of the product resource feature set is performed to obtain product hysteresis features, including:

[0121] S31, sorting the product resource feature set in time sequence to obtain a product feature sequence;

[0122] S32, performing self-attention encoding on the product feature sequence to obtain a self-attention weight;

[0123] S33, extracting the lag period of the product feature sequence according to the self-attention weight to obtain a lag coefficient;

[0124] S34, performing autocorrelation analysis on the product feature sequence according to the lag coefficient to obtain an autocorrelation coefficient;

[0125] S35. Generate product hysteresis characteristics according to the autocorrelation coefficient.

[0126] In detail, the time series sorting refers to the time series sorting of features according to the dates marked on each product resource feature in the product resource feature set. The self-attention mechanism can be used for attention encoding to obtain self-attention weights, which are the attention weights of each product feature in the product feature sequence.

[0127] Specifically, the lag period extraction refers to screening out a number of feature numbers with larger weight values ​​in the self-attention weights, and calculating the lag coefficient according to the difference between the numbers.

[0128] In detail, the following autocorrelation algorithm is used to perform autocorrelation analysis on the product feature sequence according to the lag coefficient to obtain the autocorrelation coefficient: in, The hysteresis coefficient is The autocorrelation coefficient when refers to the hysteresis coefficient, is the feature number, is the total number of features of the product feature sequence, is the first in the product feature sequence Features, is the first in the product feature sequence Features.

[0129] Specifically, the autocorrelation algorithm can measure the correlation of the sequence itself according to the interdependence between time steps in the product feature sequence, thereby discovering the periodicity or trend association of the data, and the ACF function can be used to generate product lag characteristics according to the autocorrelation coefficient.

[0130] In the embodiment of the present invention, extracting the time series feature of the product resource feature set according to the product hysteresis feature to obtain the product time series feature includes:

[0131] generating a product feature sequence according to the product resource feature set;

[0132] Performing a stationarity check on the product feature sequence to obtain a feature stationarity coefficient;

[0133] Performing a differential operation on the product feature sequence according to the feature stability coefficient to obtain a differential product feature sequence;

[0134] Performing moving average analysis on the differential product feature sequence to obtain partial regression features;

[0135] The time series characteristics of the differential product feature sequence are fitted according to the product hysteresis characteristics and the partial regression characteristics to obtain product time series characteristics.

[0136] In detail, the method of generating a product feature sequence based on the product resource feature set is consistent with the method of performing time-sequential sorting on the product resource feature set to obtain a product feature sequence in the above step S3, and will not be repeated here.

[0137] Specifically, the ADF verification algorithm or the KPSS verification algorithm can be used to perform a stationarity check. The characteristic stationarity coefficient is a coefficient used to indicate whether the product feature sequence is stationary.

[0138] In detail, performing a differential operation on the product feature sequence according to the feature stability coefficient means performing a differential operation when the feature stability coefficient is less than a preset stability threshold, and the differential can be a first-order differential or a second-order differential.

[0139] Specifically, the Yule-Walker equation or the OLS regression method can be used to perform moving average analysis, and the ARIMA model can be used to fit time series characteristics.

[0140] In the embodiment of the present invention, by performing autocorrelation analysis and time series feature extraction, the periodicity or trend correlation of product data can be discovered, which facilitates the subsequent prediction and analysis of product data in future time periods and improves the accuracy of sales data analysis.

[0141] S4. Forward propagation is performed on the product time series features to obtain analysis product data, and unsaleable product inventory is extracted from the analysis product data.

[0142] In detail, forward propagation can be performed using a long short-term memory network (LSTM) or a vector autoregressive integrated moving average model (VARIMA), whose input is product time series features and whose output is product data for a future time period. The analyzed product data is supply chain data, logistics data, and inventory data for various products in a future time period. The unsaleable product inventory refers to products that are unsaleable in a future time period and the corresponding inventory quantity.

[0143] Specifically, extracting the slow-moving product inventory from the analyzed product data includes:

[0144] Splitting the analyzed product data into product data sets according to preset product categories;

[0145] Performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set;

[0146] Using a preset slow-selling threshold value to perform slow-selling statistics on the product inventory quantity set to obtain a slow-selling product quantity set;

[0147] Generate a slow-moving product inventory according to the slow-moving product quantity set and the product data set.

[0148] In detail, the product category refers to the type of product, such as the style of clothing and the brand of electrical appliances.

[0149] Specifically, the performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set includes:

[0150] Selecting product data in the product data set one by one as target product data, and extracting product supply chain data, product initial inventory data and product logistics data from the target product data;

[0151] Extracting product warehousing data from the product supply chain data and the product logistics data;

[0152] Extracting product delivery data from the product logistics data;

[0153] Calculate the product inventory quantity according to the product entry data, the product exit data and the product initial inventory data;

[0154] The product inventory quantities of all target product data in the product data set are aggregated into a product inventory quantity set.

[0155] In detail, the product supply chain data, product initial inventory data and product logistics data are data such as the supply chain, initial inventory and logistics of the product data. The product warehousing data is the warehousing data in the product supply chain data and the product logistics data within a specific time. The product inventory quantity is equal to the product warehousing data plus the product initial inventory data minus the product outbound data.

[0156] Specifically, the slow-selling statistics refers to taking the product inventory quantity in the product inventory quantity set that is greater than a preset slow-selling threshold as the number of slow-selling products, and aggregating all the slow-selling product quantities into a slow-selling product quantity set; and generating the slow-selling product inventory based on the slow-selling product quantity set and the product data set refers to labeling the slow-selling product quantity set with product categories according to the corresponding product tags in the product data set to obtain the slow-selling product inventory.

[0157] In the embodiment of the present invention, by performing forward propagation and extracting the inventory of slow-moving products, the changes in product inventory in future time periods can be analyzed, thereby improving the accuracy of product sales analysis, and can screen out slow-moving products, facilitating subsequent targeted sales analysis.

[0158] S5. Extract product sales characteristics corresponding to the slow-moving product inventory from the enterprise resource data, and perform sales strategy analysis on the slow-moving product inventory based on the product sales characteristics to obtain a product sales strategy.

[0159] In detail, extracting the product sales features corresponding to the slow-moving product inventory from the enterprise resource data includes:

[0160] Extracting slow-moving product category characteristics from the slow-moving product inventory;

[0161] Extracting product sales data from the enterprise resource data according to the characteristics of the slow-selling product categories;

[0162] Extracting a product user group from the product sales data, and performing feature extraction on the product user group to obtain product user features;

[0163] Performing sales cycle analysis on the product sales data to obtain sales cycle characteristics;

[0164] Extracting product price features from the product sales data;

[0165] Product sales characteristics are generated according to the product price characteristics, the product user characteristics and the sales cycle characteristics.

[0166] In detail, the slow-selling product category feature refers to the category of the corresponding slow-selling product in the slow-selling product inventory, and the product sales data refers to the sales record data of the corresponding category of products recorded in the enterprise resource data in the past time period.

[0167] Specifically, the feature extraction of the product user group refers to extracting features such as user age, gender and income, the sales cycle analysis refers to calculating the sales frequency of each time period as a sales cycle feature, and the product price feature refers to the price of the corresponding type of product.

[0168] In the embodiment of the present invention, the sales strategy analysis is performed on the slow-selling product inventory according to the product sales characteristics to obtain the product sales strategy, including:

[0169] Extracting the slow-moving product category characteristics and slow-moving quantity characteristics from the slow-moving product inventory;

[0170] Performing sales feature splitting on the product sales feature, the unsalable product category feature, and the unsalable quantity feature to obtain a sales feature set;

[0171] Performing regression analysis on the sales feature set to obtain standard sales features;

[0172] Perform strategy mapping on the standard sales features to obtain product sales strategies.

[0173] In detail, the unsaleable quantity feature refers to the feature of the unsaleable quantity corresponding to each product in the unsaleable product inventory. The sales feature splitting can be performed by using the leaf nodes of a random forest model whose input is product sales feature, unsaleable product category feature and unsaleable quantity feature and whose output is product sales strategy. The decision layer of the random forest model can be used to make strategy decisions.

[0174] Specifically, the regression analysis refers to using the linear regression model of the random forest model to perform regression feature calculations on the sales feature set, and the strategy mapping refers to mapping the corresponding sales features into specific product sales strategies. The product sales strategies may be strategies such as the extent of price reduction for the product, advertising promotion in a specific time period, or advertising promotion targeting a specific group of people.

[0175] In the embodiment of the present invention, by extracting product sales features and performing sales strategy analysis, it is possible to analyze suitable promotion strategies based on historical sales data of products that may be slow-selling and perform sales analysis, thereby improving the accuracy of sales analysis.

[0176] The present invention can perform resource inventory data analysis from multiple dimensions by performing data warehouse extraction and inventory resource feature analysis, thereby improving the comprehensiveness of data during sales analysis and thus improving the accuracy of sales analysis; by performing clustering dimensionality reduction and correlation dimensionality reduction, it can reduce the feature dimensions of inventory resource features, remove redundancy and multicollinearity between features, and improve the efficiency of big data analysis; by performing autocorrelation analysis and time series feature extraction, it can discover the periodicity or trend correlation of product data, facilitate subsequent predictive analysis of product data in future time periods, and improve the accuracy of sales data analysis.

[0177] By performing forward propagation and extracting the inventory of unsalable products, the changes in product inventory in the future time period can be analyzed, thereby improving the accuracy of product sales analysis, and unsalable products can be screened out to facilitate subsequent targeted sales analysis. By extracting product sales characteristics and performing sales strategy analysis, suitable promotion strategies can be analyzed based on the historical sales data of products that may be unsalable for sales analysis, thereby improving the accuracy of sales analysis. Therefore, the sales big data analysis method based on cross-border e-commerce ERP proposed in the present invention can solve the problem of low accuracy when performing sales analysis.

[0178] like Figure 4 As shown, it is a functional module diagram of a sales big data analysis system based on cross-border e-commerce ERP provided by one embodiment of the present invention.

[0179] The sales big data analysis system 100 based on cross-border e-commerce ERP of the present invention can be installed in an electronic device. According to the functions implemented, the sales big data analysis system 100 based on cross-border e-commerce ERP can include a feature extraction module 101, a correlation dimension reduction module 102, a time series analysis module 103, a slow-selling analysis module 104 and a sales analysis module 105. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0180] In this embodiment, the functions of each module / unit are as follows:

[0181] The feature extraction module 101 is used to extract the data warehouse of the enterprise management platform to obtain enterprise resource data, and extract the inventory resource feature set from the enterprise resource data;

[0182] The correlation dimension reduction module 102 is used to perform clustering dimension reduction and correlation dimension reduction on the inventory resource feature set to obtain a product resource feature set, wherein the clustering dimension reduction and correlation dimension reduction on the inventory resource feature set to obtain the product resource feature set includes: performing central feature screening on the inventory resource feature set according to Gaussian distribution to obtain a primary central resource feature group; performing neighborhood division on the primary central resource feature group according to the inventory resource feature set to obtain a primary resource feature group set; selecting primary resource feature groups in the primary resource feature group set as target primary resource feature groups one by one, and using the primary central resource features corresponding to the target primary resource feature groups in the primary central resource feature group as target primary central resource features; selecting inventory resource features in the inventory resource feature set one by one as target inventory resource features, and calculating the product feature distance between the target inventory resource features and the target primary resource feature group according to the target primary central resource features using the following product feature distance algorithm: in, refers to the characteristic distance of the product, is the maximum value symbol, is the feature number, is the total number of features of the target primary resource feature group, is the first in the target primary resource feature group The weight of the inventory resource characteristics, is the first in the target primary resource feature group Inventory resource characteristics, is the target inventory resource characteristic, is the target inventory resource characteristic The weight of is the inner product symbol, is the modulo symbol, is the target primary center resource characteristic, is the weight of the target primary central resource feature; generating a product feature distance set according to the product feature distances of all target inventory resource features in the inventory resource feature set, and updating the target primary resource feature group into a target secondary resource feature group according to the product feature distance set; generating a secondary resource feature group set according to all target secondary resource feature groups, iteratively updating the secondary resource feature group set to obtain a product resource feature set; performing correlation dimension reduction on the product resource feature set to obtain a product resource feature set;

[0183] The time series analysis module 103 is used to perform autocorrelation analysis on the product resource feature set to obtain product hysteresis features, and extract time series features from the product resource feature set according to the product hysteresis features to obtain product time series features;

[0184] The slow-selling analysis module 104 is used to forward-propagate the product time series features to obtain analysis product data, and extract the slow-selling product inventory from the analysis product data;

[0185] The sales analysis module 105 is used to extract the product sales characteristics corresponding to the slow-moving product inventory from the enterprise resource data, and perform sales strategy analysis on the slow-moving product inventory based on the product sales characteristics to obtain a product sales strategy.

[0186] In detail, each module described in the sales big data analysis system 100 based on cross-border e-commerce ERP in the embodiment of the present invention is used in the same manner as described above. Figure 1 The sales big data analysis method based on cross-border e-commerce ERP described in the article uses the same technical means and can produce the same technical effects, so I will not go into details here.

[0187] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0188] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0189] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0190] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0191] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0192] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0193] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or systems stated in the system embodiment can also be implemented by one unit or system through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any specific order.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A sales big data analysis method based on cross-border e-commerce ERP, characterized in that: The method comprises: Extracting data from the enterprise management platform to obtain enterprise resource data, and extracting inventory resource feature sets from the enterprise resource data; The inventory resource feature set is clustered and reduced in dimension and correlation to obtain a product resource feature set, wherein the inventory resource feature set is clustered and reduced in dimension and correlation to obtain a product resource feature set, including: performing central feature screening on the inventory resource feature set according to Gaussian distribution to obtain a primary central resource feature group; performing neighborhood division on the primary central resource feature group according to the inventory resource feature set to obtain a primary resource feature group set; selecting primary resource feature groups in the primary resource feature group set one by one as target primary resource feature groups, and using the primary central resource features corresponding to the target primary resource feature groups in the primary central resource feature group as target primary central resource features; selecting inventory resource features in the inventory resource feature set one by one as target inventory resource features, and calculating the product feature distance between the target inventory resource features and the target primary resource feature group according to the target primary central resource features using the following product feature distance algorithm: Where L is the product feature distance, max is the maximum value symbol, n is the feature number, N is the total number of features in the target primary resource feature group, q n is the weight of the nth inventory resource feature in the target primary resource feature group, α n is the nth inventory resource feature in the target primary resource feature group, β is the target inventory resource feature, q β is the weight of the target inventory resource feature β, · is the inner product symbol, | | is the modulus symbol, γ is the target primary central resource feature, q γ is the weight of the target primary central resource feature; generating a product feature distance set according to the product feature distances of all target inventory resource features in the inventory resource feature set, and updating the target primary resource feature group into a target secondary resource feature group according to the product feature distance set; generating a secondary resource feature group set according to all target secondary resource feature groups, iteratively updating the secondary resource feature group set to obtain a product resource feature set; performing correlation dimension reduction on the product resource feature set to obtain a product resource feature set; Performing autocorrelation analysis on the product resource feature set to obtain product hysteresis features, and performing time series feature extraction on the product resource feature set according to the product hysteresis features to obtain product time series features; Forward propagation is performed on the product time series features to obtain analysis product data, and unsaleable product inventory is extracted from the analysis product data, wherein the analysis product data refers to the supply chain data, logistics data and inventory data of various products in a future time period, and the unsaleable product inventory refers to the products that are unsaleable in a future time period and the corresponding inventory quantity. The extraction of unsaleable product inventory from the analysis product data includes: splitting the analysis product data into product data sets according to preset product categories; performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set; performing unsale statistics on the product inventory quantity set using a preset unsale threshold to obtain a unsaleable product quantity set; generating unsaleable product inventory according to the unsaleable product quantity set and the product data set; The product sales characteristics corresponding to the slow-moving product inventory are extracted from the enterprise resource data, and a sales strategy analysis is performed on the slow-moving product inventory based on the product sales characteristics to obtain a product sales strategy.

2. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1 is characterized in that: The step of extracting the inventory resource feature set from the enterprise resource data includes: Extracting data attributes from the enterprise resource data to obtain a resource data attribute set; Performing inventory keyword matching on the resource data attribute set to obtain an inventory resource attribute set; extracting an inventory resource data set from the enterprise resource data according to the inventory resource attribute set; Performing data standardization and date marking operations on the inventory resource data set to obtain a standard inventory data set; Feature extraction and feature weighting are performed on the standard inventory data set to obtain an inventory resource feature set.

3. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1 is characterized in that: The relevance dimension reduction is performed on the product resource feature set to obtain a product resource feature set, including: Performing feature standardization operations on each product resource feature class in the product resource feature class set to obtain a standard product feature class set; The product correlation between each standard product feature class in the standard product feature class set is calculated using the following product correlation algorithm: Among them, p(C i ,C j ) refers to the standard product feature class C in the standard product feature class set i and standard product feature class C j product relevance, i, j are the serial numbers of the standard product feature classes in the standard product feature class set, m i It is the standard product feature class C i The total number of features, m j It is the standard product feature class C j The total number of features, k and l are the feature numbers, cov() is the covariance operator, σ() is the standard deviation operator, x ik Refers to the standard product feature class C i The kth feature in x jl Refers to the standard product feature class C j The lth feature in ; Filtering out relevant feature group sets from the standard product feature set according to the product relevance; Merging related features of the related feature cluster group set to obtain a merged feature cluster set; updating the standard product feature set into a decorrelated feature set according to the merged feature set; The principal component dimension reduction is performed on each decorrelation feature class in the decorrelation feature class set to obtain a product resource feature set.

4. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1 is characterized in that: The autocorrelation analysis of the product resource feature set is performed to obtain product hysteresis features, including: Sorting the product resource feature set in time sequence to obtain a product feature sequence; Performing self-attention encoding on the product feature sequence to obtain a self-attention weight; Extracting the lag period of the product feature sequence according to the self-attention weight to obtain a lag coefficient; The following autocorrelation algorithm is used to perform autocorrelation analysis on the product feature sequence according to the lag coefficient to obtain the autocorrelation coefficient: Among them, ρ r is the autocorrelation coefficient when the lag coefficient is r, r refers to the lag coefficient, t is the feature number, T is the total number of features in the product feature sequence, y t is the tth feature in the product feature sequence, y t+r is the t+rth feature in the product feature sequence; A product lag characteristic is generated based on the autocorrelation coefficient.

5. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1, characterized in that: The extracting the time series feature of the product resource feature set according to the product hysteresis feature to obtain the product time series feature includes: generating a product feature sequence according to the product resource feature set; Performing a stationarity check on the product feature sequence to obtain a feature stationarity coefficient; Performing a differential operation on the product feature sequence according to the feature stability coefficient to obtain a differential product feature sequence; Performing moving average analysis on the differential product feature sequence to obtain partial regression features; The time series characteristics of the differential product feature sequence are fitted according to the product hysteresis characteristics and the partial regression characteristics to obtain product time series characteristics.

6. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1, characterized in that: The performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set includes: Selecting product data in the product data set one by one as target product data, and extracting product supply chain data, product initial inventory data and product logistics data from the target product data; Extracting product warehousing data from the product supply chain data and the product logistics data; Extracting product delivery data from the product logistics data; Calculate the product inventory quantity according to the product entry data, the product exit data and the product initial inventory data; The product inventory quantities of all target product data in the product data set are aggregated into a product inventory quantity set.

7. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1, characterized in that: The step of extracting the product sales features corresponding to the slow-moving product inventory from the enterprise resource data includes: Extracting slow-moving product category characteristics from the slow-moving product inventory; Extracting product sales data from the enterprise resource data according to the characteristics of the slow-selling product categories; Extracting a product user group from the product sales data, and performing feature extraction on the product user group to obtain product user features; Performing sales cycle analysis on the product sales data to obtain sales cycle characteristics; Extracting product price features from the product sales data; Product sales characteristics are generated according to the product price characteristics, the product user characteristics and the sales cycle characteristics.

8. The sales big data analysis method based on cross-border e-commerce ERP as claimed in claim 1, characterized in that: The performing of sales strategy analysis on the slow-moving product inventory according to the product sales characteristics to obtain a product sales strategy includes: Extracting the slow-moving product category characteristics and slow-moving quantity characteristics from the slow-moving product inventory; Performing sales feature splitting on the product sales feature, the unsalable product category feature, and the unsalable quantity feature to obtain a sales feature set; Performing regression analysis on the sales feature set to obtain standard sales features; Perform strategy mapping on the standard sales features to obtain product sales strategies.

9. A sales big data analysis system based on cross-border e-commerce ERP, characterized in that: The system comprises: A feature extraction module is used to extract data from the enterprise management platform, obtain enterprise resource data, and extract inventory resource feature sets from the enterprise resource data; The correlation dimension reduction module is used to perform clustering dimension reduction and correlation dimension reduction on the inventory resource feature set to obtain a product resource feature set, wherein the clustering dimension reduction and correlation dimension reduction on the inventory resource feature set to obtain the product resource feature set includes: performing central feature screening on the inventory resource feature set according to Gaussian distribution to obtain a primary central resource feature group; performing neighborhood division on the primary central resource feature group according to the inventory resource feature set to obtain a primary resource feature group set; selecting primary resource feature groups in the primary resource feature group set as target primary resource feature groups one by one, and using the primary central resource features corresponding to the target primary resource feature groups in the primary central resource feature group as target primary central resource features; selecting inventory resource features in the inventory resource feature set as target inventory resource features one by one, and calculating the product feature distance between the target inventory resource features and the target primary resource feature group according to the target primary central resource features using the following product feature distance algorithm: Where L is the product feature distance, max is the maximum value symbol, n is the feature number, N is the total number of features in the target primary resource feature group, q n is the weight of the nth inventory resource feature in the target primary resource feature group, α n is the nth inventory resource feature in the target primary resource feature group, β is the target inventory resource feature, q β is the weight of the target inventory resource feature β, · is the inner product symbol, | | is the modulus symbol, γ is the target primary central resource feature, q γ is the weight of the target primary central resource feature; generating a product feature distance set according to the product feature distances of all target inventory resource features in the inventory resource feature set, and updating the target primary resource feature group into a target secondary resource feature group according to the product feature distance set; generating a secondary resource feature group set according to all target secondary resource feature groups, iteratively updating the secondary resource feature group set to obtain a product resource feature set; performing correlation dimension reduction on the product resource feature set to obtain a product resource feature set; A time series analysis module, used for performing autocorrelation analysis on the product resource feature set to obtain product hysteresis features, and performing time series feature extraction on the product resource feature set according to the product hysteresis features to obtain product time series features; The slow-selling analysis module is used to forward-propagate the product time series features to obtain analysis product data, and extract the slow-selling product inventory from the analysis product data, wherein the analysis product data refers to the supply chain data, logistics data and inventory data of various products in a future time period, and the slow-selling product inventory refers to the products that are slow-selling in a future time period and the corresponding inventory quantity. The extraction of the slow-selling product inventory from the analysis product data includes: splitting the analysis product data into product data sets according to preset product categories; performing inventory analysis on each product data in the product data set to obtain a product inventory quantity set; performing slow-selling statistics on the product inventory quantity set using a preset slow-selling threshold to obtain a slow-selling product quantity set; generating a slow-selling product inventory according to the slow-selling product quantity set and the product data set; The sales analysis module is used to extract the product sales characteristics corresponding to the slow-moving product inventory from the enterprise resource data, and perform sales strategy analysis on the slow-moving product inventory based on the product sales characteristics to obtain a product sales strategy.

Images