A data value visualization analysis system

Abstract

The present invention discloses a data value visualization analysis system, which relates to the field of data analysis technology, including: establishing an operation database of a target supermarket; analyzing the sales data change trend in the operation database, and evaluating the commodity sales status of the target supermarket; performing correlation analysis based on the commodity sales status of the target supermarket and the customer data in the operation database, and building a customer behavior portrait database; dynamically allocating commodity thermal values ​​for inventory data in historical operation data based on the customer behavior portrait database and the commodity sales status of the target supermarket, and obtaining the dynamic thermal vector index of the inventory commodities of the target supermarket; adjusting the operation strategy plan of the target supermarket based on the dynamic thermal vector index of the inventory commodities of the target supermarket. The advantages of the present invention are: improving the operation efficiency of supermarkets and realizing precision marketing.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a data value visualization analysis system. Background Art

[0002] Data value analysis is a process of in-depth mining, sorting, analyzing and interpreting the collected data to reveal the hidden information behind the data, gain insights into business trends, and evaluate the economic value and social impact of data assets. It aims to help organizations or individuals understand the true meaning of data, so as to make more accurate and effective decisions, optimize resource allocation, and improve business efficiency and competitiveness.

[0003] At present, most of the operating models of large-scale commercial supermarkets in China are comprehensive O2O models. The advantage is that they can meet the needs of most customers. The disadvantage is that the depth and breadth of data collection and utilization are insufficient, and there is a lack of comprehensive customer behavior analysis, accurate market trend forecasting, and effective inventory management optimization, resulting in the failure to fully convert data into improving business decision-making efficiency. Summary of the invention

[0004] In order to solve the above technical problems, a data value visualization analysis system is provided. This technical solution solves the problem that most of the business models of large domestic commercial supermarkets are comprehensive O2O models. The advantage is that they can meet the needs of most customers, but the disadvantage is that the depth and breadth of data collection and utilization are insufficient, and there is a lack of comprehensive customer behavior analysis, accurate market trend forecasting, and effective inventory management optimization, resulting in the failure to fully convert data into improving business decision-making efficiency.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A data value visualization analysis system, comprising:

[0007] Data collection module, which is used to obtain the historical operation data of the large commercial supermarket to be analyzed and establish the operation database of the target supermarket; the historical operation data includes: sales data, inventory data, and customer information;

[0008] A sales status analysis module, the sales status analysis module is electrically connected to the data acquisition module, and is used to analyze the sales data change trend in the operation database based on the operation database of the target supermarket, and evaluate the sales status of the goods of the target supermarket;

[0009] A behavior status analysis module, which is electrically connected to the sales status analysis module. The behavior status analysis module is used to perform correlation analysis based on the sales status of the goods in the target supermarket and the customer data in the operation database to build a customer behavior portrait database;

[0010] A thermal state analysis module, the thermal state analysis module is electrically connected to the behavior state analysis module and the sales state analysis module, and the thermal state analysis module is used to dynamically allocate thermal values ​​of commodities for inventory data in historical operation data based on a customer behavior portrait database and commodity sales status of a target supermarket, and obtain dynamic thermal vector indicators of inventory commodities of the target supermarket;

[0011] The strategy adjustment module is electrically connected to the thermal state analysis module, and is used to adjust the operation strategy plan of the target supermarket based on the dynamic thermal vector index of the inventory goods of the target supermarket.

[0012] Preferably, the sales status analysis module includes:

[0013] The time series data unit, based on the sales data in the operation database, marks the sales timestamps of the sales commodities in the sales data to form the sales commodity time series data;

[0014] A seasonal time series data unit is based on the sales commodity time series data, and divides the sales commodity time series data of the same type according to the seasonal cycle to obtain the sales commodity seasonal time series data;

[0015] A seasonality calculation unit, based on the seasonal time series data of sales commodities, calculates the average sales volume of sales commodities in the corresponding month of each season in the seasonal time series data of sales commodities and normalizes the data to obtain the seasonality index of sales commodities;

[0016] A seasonal trend calculation unit calculates the seasonal trend component coefficient of the sales commodity based on the seasonal index of the sales commodity and the seasonal time series data of the sales commodity;

[0017] A model building unit, which builds a linear analysis model of the commodity sales status of the target supermarket;

[0018] The model evaluation unit uses the seasonal trend component coefficient of the sales commodity as the influencing factor of the sales commodity in the sales data, substitutes the seasonal time series data of the sales commodity into the linear analysis model of the commodity sales status of the target supermarket, and evaluates the commodity sales status index of the target supermarket;

[0019] Among them, the seasonal trend component coefficient of sales commodities is calculated as follows:

[0020]

[0021] In the formula, ρ i is the seasonal trend component coefficient of the i-th sales product, S ij is the seasonal index of the i-th sales product in the j-th season, is the average sales volume of the i-th sales product in the j-th season in the seasonal time series data of sales products, v j is the total number of months in season j, a ij (k) is the sales value of the i-th sales product in the k-th month of the j-th season in the seasonal time series data of sales products.

[0022] Preferably, the behavior state analysis module includes:

[0023] The numbering unit binds a unique ID number to each customer based on the target supermarket's historical customer consumption records;

[0024] The standardization unit obtains the consumption records of historical customers of the target supermarket and performs standardization processing to obtain standardized data of the consumption records of historical customers;

[0025] The feature extraction unit, based on the standardized data of historical customer consumption records, marks the consumption characteristics of each customer and obtains the consumption characteristic preference data of historical customers;

[0026] The clustering unit, based on the consumption characteristic preference data of historical customers, performs fuzzy clustering of the same characteristics on the consumption characteristic preference data of historical customers according to the month index under each season, and obtains several consumption preference characteristic data sets of historical customers;

[0027] The user portrait unit counts the multi-dimensional label data contribution value corresponding to the consumption preference feature data cluster in the set of several consumption preference feature data clusters of each historical customer, determines the preference behavior label of the historical customer, and builds the customer behavior portrait database;

[0028] Preferably, the same feature fuzzy clustering is performed on the consumption feature preference data of historical customers according to the month in each season as follows:

[0029]

[0030] In the formula, f m (U,v) is the clustering function, where U is the membership matrix, v is the cluster center matrix, and x i'j' (k) is the membership degree of the j'th consumption characteristic preference data under the k'th month index to the i'th cluster, v i' is the cluster center of the i'th cluster, c is the total number of clusters, n is the total number of consumption characteristic preference data, and m is the fuzzy factor.

[0031] Preferably, the thermal state analysis module internally includes:

[0032] The second model unit builds a sales change trend model of the target supermarket based on AR autoregression;

[0033] The influence coefficient unit is based on the customer behavior portrait database and the product sales status index of the target supermarket, and statistically analyzes the linear influence trend index between the behavior portraits of all customers under the monthly index and the product sales status index of the target supermarket;

[0034] The second model analysis unit uses the linear impact trend index between the behavior profiles of all customers per unit time and the commodity sales status index of the target supermarket as the sales impact factor, substitutes the sales data in the operation database of the target supermarket into the commodity sales transformation trend model of the target supermarket, and predicts the future commodity sales volume of the target supermarket;

[0035] The heat distribution unit calculates the dynamic heat vector index of the target supermarket's inventory goods based on the customer's behavior profile, the predicted future sales volume of the target supermarket's goods, and the target supermarket's inventory quantity;

[0036] Preferably, the commodity sales change trend model of the target supermarket is specifically:

[0037]

[0038] Sales t+1 is the sales volume of the target supermarket at the t+1th unit time order, Sales t-z+1 is the historical sales data under t-z+1 unit time, β is a constant term, ∈ z+1 is the error term at the z+1th unit time, φ z is the z-th order autoregressive coefficient, q is the total number of learning orders, and D is the linear impact trend index between the behavior profiles of all customers and the commodity sales status index of the target supermarket;

[0039] The dynamic heat vector index of the inventory of the target supermarket is calculated as follows:

[0040]

[0041] In the formula, H τ is the dynamic heat vector index of the τth inventory item in the target supermarket, Quantum t is the inventory quantity of the target supermarket at the tth unit time, and γ1 and γ2 are weight coefficients.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention proposes a data value visualization analysis system which establishes a behavioral portrait of consumer customers, analyzes customer preferences, assigns dynamic thermal values ​​to inventory goods, and adjusts the purchase quantity of target supermarkets and changes the layout of thermal goods based on the dynamic thermal values ​​assigned to inventory goods. The advantages of the present invention are: improving the operating efficiency of supermarkets and achieving precision marketing. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a framework diagram of a data value visualization analysis system;

[0045] Figure 2 This is the internal framework diagram of the sales status analysis module;

[0046] Figure 3 This is the internal framework diagram of the behavior state analysis module;

[0047] Figure 4 This is the internal framework diagram of the behavior state analysis module. DETAILED DESCRIPTION

[0048] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0049] Reference Figure 1 As shown, a data value visualization analysis system includes:

[0050] Data collection module, which is used to obtain the historical operation data of the large commercial supermarket to be analyzed and establish the operation database of the target supermarket; the historical operation data includes: sales data, inventory data, and customer information;

[0051] A sales status analysis module, the sales status analysis module is electrically connected to the data acquisition module, and is used to analyze the sales data change trend in the operation database based on the operation database of the target supermarket, and evaluate the sales status of the goods of the target supermarket;

[0052] A behavior status analysis module, which is electrically connected to the sales status analysis module. The behavior status analysis module is used to perform correlation analysis based on the sales status of the goods in the target supermarket and the customer data in the operation database to build a customer behavior portrait database;

[0053] A thermal state analysis module, the thermal state analysis module is electrically connected to the behavior state analysis module and the sales state analysis module, and the thermal state analysis module is used to dynamically allocate thermal values ​​of commodities for inventory data in historical operation data based on a customer behavior portrait database and commodity sales status of a target supermarket, and obtain dynamic thermal vector indicators of inventory commodities of the target supermarket;

[0054] The strategy adjustment module is electrically connected to the thermal state analysis module, and is used to adjust the operation strategy plan of the target supermarket based on the dynamic thermal vector index of the inventory goods of the target supermarket.

[0055] It is understandable that due to the attributes of the goods and external factors, the sales status of goods in commercial supermarkets usually presents a normal distribution at a high latitude. Therefore, by establishing a behavioral portrait of consumer customers and analyzing customer preferences, dynamic thermal values ​​are assigned to inventory goods, and based on the dynamic thermal values ​​assigned to inventory goods, the purchase quantity of the target supermarket is adjusted and the layout of thermal goods is changed. The advantages of the present invention are: improving the operating efficiency of supermarkets and realizing precision marketing;

[0056] Reference Figure 2 As shown, the sales status analysis module includes:

[0057] The time series data unit, based on the sales data in the operation database, marks the sales timestamps of the sales commodities in the sales data to form the sales commodity time series data;

[0058] A seasonal time series data unit is based on the sales commodity time series data, and divides the sales commodity time series data of the same type according to the seasonal cycle to obtain the sales commodity seasonal time series data;

[0059] A seasonality calculation unit, based on the seasonal time series data of sales commodities, calculates the average sales volume of sales commodities in the corresponding month of each season in the seasonal time series data of sales commodities and normalizes the data to obtain the seasonality index of sales commodities;

[0060] A seasonal trend calculation unit calculates the seasonal trend component coefficient of the sales commodity based on the seasonal index of the sales commodity and the seasonal time series data of the sales commodity;

[0061] A model building unit, which builds a linear analysis model of the commodity sales status of the target supermarket;

[0062] The model evaluation unit uses the seasonal trend component coefficient of the sales commodity as the influencing factor of the sales commodity in the sales data, substitutes the seasonal time series data of the sales commodity into the linear analysis model of the commodity sales status of the target supermarket, and evaluates the commodity sales status index of the target supermarket;

[0063] Among them, the seasonal trend component coefficient of sales commodities is calculated as follows:

[0064]

[0065] In the formula, ρ i is the seasonal trend component coefficient of the i-th sales product, S ij is the seasonal index of the i-th sales product in the j-th season, is the average sales volume of the i-th sales product in the j-th season in the seasonal time series data of sales products, v j is the total number of months in season j, a ij (k) is the sales value of the i-th sales product in the k-th month of the j-th season in the seasonal time series data of sales products.

[0066] This solution can not only help supermarket managers clearly identify the seasonal changes in product sales, but also quantify the specific impact of such changes on sales status, thereby providing them with more accurate and scientific decision-making support. By evaluating the product sales status index, supermarkets can timely adjust product structure, promotion strategies, etc. to respond to market changes, optimize resource allocation, and improve operating efficiency.

[0067] Reference Figure 3 As shown, the behavior status analysis module includes:

[0068] The numbering unit binds a unique ID number to each customer based on the target supermarket's historical customer consumption records;

[0069] The standardization unit obtains the consumption records of historical customers of the target supermarket and performs standardization processing to obtain standardized data of the consumption records of historical customers;

[0070] The feature extraction unit marks the consumption characteristics of each customer based on the standardized data of the historical customer's consumption records, and obtains the consumption characteristic preference data of the historical customers; the consumption characteristics include: the category, brand, price range, purchase frequency, and purchase time of the purchased goods;

[0071] The clustering unit, based on the consumption characteristic preference data of historical customers, performs fuzzy clustering of the same characteristics on the consumption characteristic preference data of historical customers according to the month index under each season, and obtains several consumption preference characteristic data sets of historical customers;

[0072] The user portrait unit counts the multi-dimensional label data contribution value corresponding to the consumption preference feature data cluster in the set of several consumption preference feature data clusters of each historical customer, determines the preference behavior label of the historical customer, and builds a customer behavior portrait database; the multi-dimensional label data includes: high consumption ability, fresh food preference, high-frequency purchaser, and active user;

[0073] Among them, the same feature fuzzy clustering is performed on the consumption feature preference data of historical customers according to the month in each season as follows:

[0074]

[0075] In the formula, f m(U,v) is the clustering function, where U is the membership matrix, v is the cluster center matrix, and x i'j' (k) is the membership degree of the j'th consumption characteristic preference data under the k'th month index to the i'th cluster, v i' is the cluster center of the i'th cluster, c is the total number of clusters, n is the total number of consumption characteristic preference data, and m is the fuzzy factor.

[0076] The implementation principle of this plan is to collect the historical customer consumption records of the target supermarkets and perform standardization processing to accurately extract the customer's consumption characteristic preferences, and use the fuzzy clustering method to cluster the consumption characteristics by season and month to identify multiple sets of customer consumption preference characteristics; by counting the contribution values ​​of these feature sets to multi-dimensional labels, each customer is accurately labeled with preference behavior, and a customer behavior portrait database is constructed, which can accurately depict the customer's consumption behavior patterns and provide supermarkets with a basis for personalized marketing and service strategies.

[0077] Reference Figure 4 As shown in the figure, the thermal state analysis module includes:

[0078] The second model unit builds a sales change trend model of the target supermarket based on AR autoregression;

[0079] The influence coefficient unit is based on the customer behavior portrait database and the product sales status index of the target supermarket, and statistically analyzes the linear influence trend index between the behavior portraits of all customers under the monthly index and the product sales status index of the target supermarket;

[0080] The second model analysis unit uses the linear impact trend index between the behavior profiles of all customers per unit time and the commodity sales status index of the target supermarket as the sales impact factor, substitutes the sales data in the operation database of the target supermarket into the commodity sales transformation trend model of the target supermarket, and predicts the future commodity sales volume of the target supermarket;

[0081] The heat distribution unit calculates the dynamic heat vector index of the target supermarket's inventory goods based on the customer's behavior profile, the predicted future sales volume of the target supermarket's goods, and the target supermarket's inventory quantity;

[0082] The specific sales change trend model of the target supermarket is:

[0083]

[0084] Sales t+1 is the sales volume of the target supermarket at the t+1th unit time order, Sales t-z+1 is the historical sales data under t-z+1 unit time, β is a constant term, ∈z+1 is the error term at the z+1th unit time, φ z is the z-th order autoregressive coefficient, q is the total number of learning orders, and D is the linear impact trend index between the behavior profiles of all customers and the commodity sales status index of the target supermarket;

[0085] The dynamic heat vector index of the inventory of the target supermarket is calculated as follows:

[0086]

[0087] In the formula, H τ is the dynamic heat vector index of the τth inventory item in the target supermarket, Quantum t is the inventory quantity of the target supermarket at the tth unit time, and γ1 and γ2 are weight coefficients.

[0088] It should be noted that the weight coefficient is used to adjust the importance of different factors in the calculation of thermal value. The weight can be set and adjusted according to the actual operation situation and goals of the supermarket, and will not be described in detail here;

[0089] This solution combines the AP autoregressive model, customer behavior portrait analysis and sales data forecasting to accurately grasp the sales trend of target supermarket products and effectively predict future sales quantities, helping supermarkets to adjust inventory strategies in advance and reduce inventory backlogs and out-of-stock risks. It can also optimize product structure according to customer preferences and market trends, improve sales efficiency and customer satisfaction, and the introduction of dynamic heat vector indicators provides supermarkets with an intuitive and quantitative way to evaluate the comprehensive performance and future potential of inventory products, providing strong support for decision-making.

[0090] The usage process of a data value visualization analysis system is as follows:

[0091] Step 1: Obtain the historical operation data of the large commercial supermarket to be analyzed and establish the operation database of the target supermarket;

[0092] Step 2: Based on the sales data in the operation database, mark the sales timestamps of the sales products in the sales data to form the sales product time series data;

[0093] Step 3: Based on the sales commodity time series data, divide the sales commodity time series data of the same type according to the seasonal cycle to obtain the sales commodity seasonal time series data;

[0094] Step 4: Based on the seasonal time series data of sales commodities, the average sales volume of the sales commodities in the corresponding month of each season in the seasonal time series data of sales commodities is calculated and normalized to obtain the seasonal index of sales commodities;

[0095] Step 5: Based on the seasonal index of sales products and the seasonal time series data of sales products, calculate the seasonal trend component coefficient of sales products;

[0096] Step 6: Establish a linear analysis model of the product sales status of the target supermarket;

[0097] Step 7: Use the seasonal trend component coefficient of the sales product as the influencing factor of the sales product in the sales data, substitute the seasonal time series data of the sales product into the linear analysis model of the product sales status of the target supermarket, and evaluate the product sales status index of the target supermarket;

[0098] Step 8: Based on the historical customer consumption records of the target supermarket, bind a unique ID number to each customer;

[0099] Step 9: Obtain the consumption records of historical customers of the target supermarket and perform standardized processing to obtain standardized data of the consumption records of historical customers;

[0100] Step 10: Based on the standardized data of historical customers’ consumption records, mark the consumption characteristics of each customer and obtain the consumption characteristic preference data of historical customers;

[0101] Step 11: Based on the consumption characteristic preference data of historical customers, perform fuzzy clustering of the same characteristics on the consumption characteristic preference data of historical customers according to the month index under each season to obtain several consumption preference characteristic data sets of historical customers;

[0102] Step 12: Count the multi-dimensional label data contribution values ​​corresponding to the consumption preference feature data clusters in the consumption preference feature data cluster sets of each historical customer, determine the preference behavior labels of the historical customers, and build a customer behavior portrait database;

[0103] Step 13: Based on AR autoregression, build a sales change trend model of the target supermarket;

[0104] Step 14: Based on the customer behavior portrait database and the product sales status index of the target supermarket, statistically analyze the linear impact trend index between the behavior portraits of all customers under the month index and the product sales status index of the target supermarket;

[0105] Step 15: Taking the linear impact trend index between the behavior profiles of all customers per unit time and the product sales status index of the target supermarket as the sales impact factor, substitute the sales data in the operation database of the target supermarket into the product sales transformation trend model of the target supermarket to predict the future product sales volume of the target supermarket;

[0106] Step 16: Based on the customer's behavior profile, the target supermarket's predicted future sales volume and inventory volume are calculated to calculate the dynamic heat vector index of the target supermarket's inventory.

[0107] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A data value visualization analysis system, characterized in that: include: Data collection module: The data collection module is used to obtain the historical operation data of the large commercial supermarket to be analyzed and establish the operation database of the target supermarket; The historical operation data includes: sales data, inventory data, and customer information; A sales status analysis module, the sales status analysis module is electrically connected to the data acquisition module, and is used to analyze the sales data change trend in the operation database based on the operation database of the target supermarket, and evaluate the sales status of the goods of the target supermarket; A behavior status analysis module, which is electrically connected to the sales status analysis module. The behavior status analysis module is used to perform correlation analysis based on the sales status of the goods in the target supermarket and the customer data in the operation database to build a customer behavior portrait database; A thermal state analysis module, the thermal state analysis module is electrically connected to the behavior state analysis module and the sales state analysis module, and the thermal state analysis module is used to dynamically allocate thermal values ​​of commodities for inventory data in historical operation data based on a customer behavior portrait database and commodity sales status of a target supermarket, and obtain dynamic thermal vector indicators of inventory commodities of the target supermarket; A strategy adjustment module, the strategy adjustment module is electrically connected to the thermal state analysis module, and the strategy adjustment module is used to adjust the operation strategy plan of the target supermarket based on the dynamic thermal vector index of the inventory goods of the target supermarket; The sales status analysis module includes: The time series data unit, based on the sales data in the operation database, marks the sales timestamps of the sales commodities in the sales data to form the sales commodity time series data; A seasonal time series data unit is based on the sales commodity time series data, and divides the sales commodity time series data of the same type according to the seasonal cycle to obtain the sales commodity seasonal time series data; A seasonality calculation unit, based on the seasonal time series data of sales commodities, calculates the average sales volume of sales commodities in the corresponding month of each season in the seasonal time series data of sales commodities and normalizes the data to obtain the seasonality index of sales commodities; A seasonal trend calculation unit calculates the seasonal trend component coefficient of the sales commodity based on the seasonal index of the sales commodity and the seasonal time series data of the sales commodity; A model building unit, which builds a linear analysis model of the commodity sales status of the target supermarket; The model evaluation unit uses the seasonal trend component coefficient of the sales commodity as the influencing factor of the sales commodity in the sales data, substitutes the seasonal time series data of the sales commodity into the linear analysis model of the commodity sales status of the target supermarket, and evaluates the commodity sales status index of the target supermarket; The thermal state analysis module includes: The second model unit builds a sales change trend model of the target supermarket based on AR autoregression; The influence coefficient unit is based on the customer behavior portrait database and the product sales status index of the target supermarket, and statistically analyzes the linear influence trend index between the behavior portraits of all customers under the monthly index and the product sales status index of the target supermarket; The second model analysis unit uses the linear impact trend index between the behavior profiles of all customers per unit time and the commodity sales status index of the target supermarket as the sales impact factor, substitutes the sales data in the operation database of the target supermarket into the commodity sales transformation trend model of the target supermarket, and predicts the future commodity sales volume of the target supermarket; The heat distribution unit calculates the dynamic heat vector index of the target supermarket's inventory goods based on the customer's behavior profile, the predicted future sales volume of the target supermarket's goods, and the target supermarket's inventory quantity; The commodity sales change trend model of the target supermarket is specifically: Sales t+1 is the sales volume of the target supermarket at the t+1th unit time order, Sales t-z+1 is the historical sales data under t-z+1 unit time, β is a constant term, ∈ z+1 is the error term at the z+1th unit time, φ z is the z-th order autoregressive coefficient, q is the total number of learning orders, and D is the linear impact trend index between the behavior profiles of all customers and the commodity sales status index of the target supermarket; The dynamic heat vector index of the inventory of the target supermarket is calculated as follows: In the formula, H τ is the dynamic heat vector index of the τth inventory item in the target supermarket, Quantum t is the inventory quantity of the target supermarket at the tth unit time, and γ1 and γ2 are weight coefficients.

2. A data value visualization analysis system according to claim 1, characterized in that: The behavior status analysis module includes: The numbering unit binds a unique ID number to each customer based on the target supermarket's historical customer consumption records; The standardization unit obtains the consumption records of historical customers of the target supermarket and performs standardization processing to obtain standardized data of the consumption records of historical customers; The feature extraction unit, based on the standardized data of historical customer consumption records, marks the consumption characteristics of each customer and obtains the consumption characteristic preference data of historical customers; The clustering unit, based on the consumption characteristic preference data of historical customers, performs fuzzy clustering of the same characteristics on the consumption characteristic preference data of historical customers according to the month index under each season, and obtains several consumption preference characteristic data sets of historical customers; The user portrait unit counts the multi-dimensional label data contribution values ​​corresponding to the consumption preference feature data clusters in the set of several consumption preference feature data clusters of each historical customer, determines the preference behavior labels of historical customers, and builds a customer behavior portrait database.

Images