Local product transaction amount calculation method and system based on multi-dimensional feature analysis

Abstract

The application discloses a local product transaction amount calculation method and system based on multi-dimensional feature analysis, relates to the technical field of data processing, and comprises the following steps: collecting multi-source related data, standardizing the multi-source related data and generating a data set, extracting product attribute-related features and matching and determining attributes, and generating a local related product list; combining the data set and the list to divide commercial outlets into professional sales stores and comprehensive retail outlets; checking the transaction authenticity of the professional sales stores, counting the transaction amount, splitting local and cross-regional transaction data of the comprehensive retail outlets, constructing an entropy weight method weighted model to calculate the proportion of local products and account for the transaction amount; and after the transaction amount is summarized, the transaction amount is counted in multiple dimensions according to local industry benchmark data dimensions to generate local related product transaction amount statistical results. The system sets corresponding function modules to realize the method, and the application improves the accuracy and authenticity of transaction amount statistics, and the statistical results can provide reliable data support for regional industry development and policy making.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for calculating local product transaction volume based on multi-dimensional feature analysis. Background Technology

[0002] In the field of regional economic statistics and industrial development monitoring, the statistics on local product transaction volume are an important data support for understanding the regional industrial development trend and formulating relevant industrial policies. At present, the calculation of local product transaction volume mainly relies on the combination of enterprise self-reporting and manual sampling survey: local product operators compile their own local product transaction data and report it to the statistical agency. The statistical agency then selects some commercial outlets to conduct manual sampling surveys, and calculates the local product transaction volume in the region after verifying the reported data.

[0003] In the implementation of the aforementioned existing technologies, due to the lack of a systematic classification system for commercial outlets and the failure to accurately identify and analyze the geographical attributes of products, as well as the lack of scientific calculations based on multi-dimensional data on commercial outlets, products, and transactions, the transaction amount calculations are completed solely by relying on data self-reported by business entities and limited data from manual sampling. This results in insufficient accuracy in the final statistical results of local product transaction amounts, making it difficult to provide accurate and reliable local product transaction data support for regional industrial development and policy formulation. Summary of the Invention

[0004] To address the technical problem in existing technologies where transaction volume calculation relies solely on limited data submitted by business entities and manual sampling, resulting in insufficient accuracy of the final statistical results for local product transaction volume, this invention provides a method and system for calculating local product transaction volume based on multi-dimensional feature analysis.

[0005] The technical solution adopted in this invention is:

[0006] The first aspect of this application provides a method for calculating local product transaction volume based on multi-dimensional feature analysis, including the following steps:

[0007] Step 1: Collect basic data of commercial outlets, product attribute data, transaction behavior data, and local industry benchmark data, and perform data standardization processing to generate a standardized dataset;

[0008] Step 2: Extract the local association features of the products based on the standardized dataset, match them with the local industry benchmark data to determine whether the products have local association attributes, and generate a list of locally associated products.

[0009] Step 3: Classify commercial outlets based on the standardized dataset and the local associated product list to generate classification results for specialty stores and general retail outlets;

[0010] Step 4: Based on the standardized dataset and the local associated product list, determine the local associated sales attributes of the professional sales stores, and calculate the local associated product transaction amount of the professional sales stores that meet the local associated sales attributes according to the determination results.

[0011] Step 5: Based on the standardized dataset and the local associated product list, calculate the proportion of local associated product transaction volume of the comprehensive retail outlets using a weighted model, and calculate the local associated product transaction volume in combination with the overall transaction data of the outlets;

[0012] Step 6: Summarize the transaction volume of local related products of the professional sales stores and comprehensive retail outlets, and conduct multi-dimensional statistical analysis according to the statistical dimensions of local industry benchmark data to generate local related product transaction volume statistics results.

[0013] The second aspect of this application provides a local product transaction volume calculation system based on multi-dimensional feature analysis, which applies the aforementioned local product transaction volume calculation method based on multi-dimensional feature analysis, including:

[0014] The multi-source data acquisition and standardization module is used to collect basic data of commercial outlets, product attribute data, transaction behavior data, and local industry benchmark data, and to perform data standardization processing to generate a standardized dataset.

[0015] The product location feature analysis and determination module is used to extract the location association features of products based on the standardized dataset, match them with the local industry benchmark data to determine whether the products have location association attributes, and generate a local associated product list.

[0016] A business outlet intelligent classification module is used to classify business outlets based on the standardized dataset and the local associated product list, and generate classification results for specialty stores and general retail outlets;

[0017] The professional sales store location determination and transaction volume statistics module is used to determine the location-related sales attributes of the professional sales stores based on the standardized dataset and the local related product list, and to count the local related product transaction volume of professional sales stores that meet the location-related sales attributes according to the determination results.

[0018] A weighted calculation module for local transaction volume of comprehensive retail outlets is used to calculate the proportion of local related product transaction volume of comprehensive retail outlets based on the standardized dataset and the local related product list, and to calculate the local related product transaction volume in combination with the overall transaction data of the outlets.

[0019] The local product transaction volume multi-dimensional summary and statistics module is used to summarize the local related product transaction volume of the professional sales stores and comprehensive retail outlets, perform multi-dimensional statistical analysis according to the statistical dimensions of local industry benchmark data, and generate local related product transaction volume statistics results.

[0020] The beneficial effects of the present invention are at least one of the following:

[0021] By linking the determination of product location association attributes with the classification of business outlets, and using standardized datasets and local associated product lists as dual criteria for classifying business outlets, this approach differs from the undifferentiated general outlet classification methods of existing technologies. This ensures that the outlet classification results are no longer independent of local product transaction characteristics, but are deeply adapted to the statistical needs of local product transaction volume. This avoids the statistical bias caused by the disconnect between classification and statistical needs in existing technologies at the outlet classification level.

[0022] This solution designs differentiated local product transaction volume calculation logics for specialty stores and general retail outlets. Unlike existing technologies that use a single accounting method for all outlets, this solution combines the operating characteristics of the two types of outlets. For specialty stores, it directly calculates transaction volume based on the results of location-based determination. For general retail outlets, it calculates the proportion of local related products using a weighted model before calculating transaction volume. This ensures that the transaction volume calculation logic matches the local product transaction patterns of different types of outlets, solving the problem that the existing single accounting method cannot adapt to the operating characteristics of different outlets. This improves the accuracy of transaction volume statistics from the perspective of calculation logic. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the method flow of Embodiment 1 of the present invention;

[0024] Figure 2 This is a system architecture block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0025] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0026] Example 1

[0027] In existing technologies, the calculation of local product transaction volume mainly adopts a combination of enterprise self-declaration and manual sampling survey. The data collection process only obtains scattered declared and sampled data without standardization. The determination of product location-related attributes relies solely on the information declared by the business entity without objective feature extraction and matching basis. The classification of commercial outlets only extracts the operating characteristics of the outlets themselves without combining them with the operating characteristics of local products. This leads to a disconnect between the classification results and the statistical needs of local product transaction volume, resulting in insufficient accuracy of the statistical results and failing to provide reliable local product transaction data support for regional industrial development and policy formulation.

[0028] To address the aforementioned issues, this embodiment provides a method for calculating local product transaction volume based on multi-dimensional feature analysis, such as... Figure 1 As shown, it includes the following steps:

[0029] Step 1: Collect basic data of business outlets, product attribute data, transaction behavior data, and local industry benchmark data, and perform data standardization processing to generate a standardized dataset.

[0030] It should be noted that basic data on commercial outlets refers to the basic information on which commercial outlets conduct business activities, including their business address, business scope, registered entity, and range of business categories.

[0031] Product attribute data refers to the inherent attribute information of products sold at commercial outlets, including product name, trademark, brand, place of origin, product barcode, packaging label, and other information.

[0032] Transaction behavior data refers to all data generated by commercial outlets in conducting product transactions, including product sales volume, product price, number of transactions, number of reviews, transaction order information, buyer-related information, logistics information, etc.

[0033] Local industry benchmark data refers to benchmark reference data related to industrial development within a locality, including the local characteristic industry category catalog, local administrative region code, transaction characteristic data of various industries in the locality, and local industry statistical dimension classification standards, etc.

[0034] Data standardization refers to the unified processing of collected heterogeneous data from multiple sources, transforming them into standardized, unified, and interconnected datasets. A standardized dataset is a unified dataset formed after standardization of multi-source data, which can be directly used for subsequent feature extraction, model building, and data computation.

[0035] In the specific implementation process, basic data of commercial outlets are collected through the business registration database and the commercial outlet operation filing system; product attribute data are collected through the product barcode database and the commercial outlet product listing system; transaction behavior data are collected through the offline POS transaction system and the e-commerce platform transaction backend; and local industry benchmark data are collected through the industry development reports and statistical bulletins issued by the local industry authorities.

[0036] The collected four types of data underwent data standardization processing. First, the data types were cleaned to remove data with missing core fields. Then, redundant data was deduplicated. Unique identifiers were established for each type of data based on business outlet ID, product barcode, and transaction order number to remove duplicate data. Subsequently, heterogeneous data of different formats were standardized, converting unstructured text and image data into structured data. Finally, the four types of data after standardization were fused, and the relationships between the data types were established based on the unique identifiers. The cleaned, deduplicated, standardized, and fused data was integrated into a standardized dataset and stored in a distributed database for subsequent processing.

[0037] For example, taking the calculation of local product transaction volume in the fresh agricultural products sector of City A as an example, basic data of 80 fresh food-related commercial outlets in City A, such as their business addresses and business scope, are collected. Product attribute data such as trademarks, place of origin labels, and product barcodes of 320 fresh food products sold at these 80 outlets are also collected. Transaction behavior data such as sales volume, number of transactions, and logistics information of these 80 outlets over the past 30 days are also collected. Local industry benchmark data such as the fresh food specialty category catalog and administrative region codes of each district in City A are also collected. The above data is cleaned, removing product attribute data without product barcodes and transaction behavior data without transaction counts. A unique identifier is established for each of the 80 outlets based on the commercial outlet ID. Duplicate basic data of outlets is removed. Unstructured place of origin text on product packaging is converted into structured local code data. Finally, the association between product attribute data and transaction behavior data is established based on product barcodes. After fusion, a standardized dataset for the fresh agricultural products sector in City A is generated.

[0038] Step 2: Extract the product's location association features based on the standardized dataset, match them with the local industry benchmark data to determine whether the product has location association attributes, and generate a local associated product list.

[0039] It should be noted that the territorial association characteristic refers to the characteristics that reflect the industrial, production, and business relationship between the product and the territorial location, and is the core basis for determining whether a product is a territorially associated product.

[0040] Considering that local niche products lack trademarks and brand operators, directly matching them based solely on basic characteristics could easily lead to them being missed as products from outside the local area, resulting in incomplete product location determination. In one possible implementation, step 2 includes the following sub-steps:

[0041] Step 2.1: Extract the basic geographical association features and distinctive geographical association features of the products based on the standardized dataset. The basic geographical association features include trademarks, brands, and production location codes. The distinctive geographical association features include geographical industry identifiers, product distribution links, and production process attribute features.

[0042] Among them, basic territorial association characteristics refer to the standardized and regulated territorial association characteristics of a product, which are the basis for determining the territorial association attributes of a product; distinctive territorial association characteristics refer to the non-standardized territorial association characteristics of a product, which are mainly applicable to determining the territorial association attributes of local niche specialty products without trademarks or brands.

[0043] Local industry identification refers to the identification information that a product belongs to a local characteristic industry category; product circulation chain refers to the circulation information of the product from the production end to the sales end; production process attribute characteristics refer to the attribute information that the product is locally handmade.

[0044] Step 2.2: Perform the first matching between the basic territorial association features and the territorial industry benchmark data.

[0045] Step 2.3: For products that are not identified as having a geographical association attribute in the first match, perform a second match between the distinctive geographical association features and the geographical industry benchmark data.

[0046] Step 2.4: Based on the results of the first and second matching, determine whether the product has a local association attribute and generate a list of locally associated products.

[0047] Among them, the local association attribute refers to the attribute that the product has an industrial, production, or business relationship with the locality; the local association product list refers to a list that records information on all products that are determined to have local association attributes, including product name, product barcode, and affiliated commercial outlets.

[0048] For example, taking the fresh agricultural products sector in City A as an example, basic local association features such as trademarks, brands, and places of production of 320 fresh products were extracted from a standardized dataset, as well as distinctive local association features such as local industry identifiers, product distribution channels, and production process attributes. The basic local association features were first matched with local industry benchmark data such as the City A Fresh Specialty Category Catalog and the City A Administrative Region Code, and 210 products, such as strawberries and rice, whose production location was marked as a certain jurisdiction in City A, were determined to have local association attributes. A second match was performed on the 110 products that were not determined in the first match, and 28 niche specialty products, such as handmade tofu skin and homemade pickles from villages and towns, which belong to the City A Fresh Specialty Category Catalog and whose distribution channels are directly connected from farmers in City A to community outlets in City A, were determined to have local association attributes. Combining the results of the two matches, 238 products were compiled into a list of locally associated products in the fresh agricultural products sector of City A.

[0049] This step extracts both the basic and distinctive local characteristics of products and uses a combination of initial and secondary matching to quickly identify conventional products with standardized identifiers and accurately identify local niche products without standardized identifiers. This effectively improves the comprehensiveness of product location association attribute determination, reduces product omissions, and the generated list of locally associated products can provide accurate product determination basis for subsequent business outlet classification and transaction volume calculation.

[0050] Step 3: Classify commercial outlets based on the standardized dataset and the local associated product list to generate classification results for specialty stores and general retail outlets.

[0051] It should be noted that the classification results of specialty stores and general retail outlets will be synchronized to steps 4 and 5, serving as the classification basis for subsequent transaction volume calculations.

[0052] Considering the different types of commercial outlets, the transaction characteristics of their local products also vary significantly. Existing technologies only extract the operational characteristics of the commercial outlets themselves for classification, without considering the operational characteristics of local products. This leads to a disconnect between the classification results and the statistical needs of local product transaction volume. In one possible implementation, step 3 includes the following sub-steps:

[0053] Step 3.1: Extract the business characteristics of the business outlets based on the standardized dataset, and extract the local product business characteristics of the business outlets based on the local related product list.

[0054] Among them, the self-operating characteristics refer to the operating characteristics of the commercial outlet itself, which are characteristics that reflect the basic operating status of the commercial outlet; the local product operating characteristics refer to the characteristics of the commercial outlet in operating local related products, which are characteristics that reflect the degree of correlation between the commercial outlet and the operation of local related products.

[0055] In the specific implementation process, based on the basic data and transaction behavior data of commercial outlets in the standardized dataset, the business characteristics of each commercial outlet are extracted, including the scope of business, the number of product categories, and the average order value. Based on the local related product list, the local product business characteristics of each commercial outlet are extracted, including the number of local related products on the shelves, the sales ratio of local related products, and the business ratio of local specialty products.

[0056] Step 3.2: Construct a network classification model by integrating the self-operating characteristics and the local product operating characteristics.

[0057] Among them, the outlet classification model refers to a machine learning model that integrates the business characteristics of a commercial outlet with the business characteristics of local products, and is used to determine the type of commercial outlet.

[0058] In the specific implementation process, the extracted self-operation characteristics are fused with the local product operation characteristics to form a classification feature set of commercial outlets. The data of commercial outlets with labeled types in the locality are used as training samples to train the random forest machine learning model. The trained model is used as the outlet classification model. The input of the model is the classification feature set of commercial outlets, and the output is the classification label of professional sales stores or comprehensive retail outlets.

[0059] Step 3.3: Input the feature data of the commercial outlets into the outlet classification model to complete the classification of specialty stores and general retail outlets and generate classification results.

[0060] In the specific implementation process, the characteristic data of the business outlets to be classified, which are integrated with the local product business characteristics, are input into the outlet classification model. The model outputs the type label of each business outlet, and the type labels of all business outlets are sorted into the classification result.

[0061] For example, taking the fresh agricultural products sector in City A as an example, the business characteristics of 80 fresh food-related commercial outlets, such as their business scope and the number of product categories they operate, are extracted from a standardized dataset. Local product characteristics, such as the number of locally related products on the shelves and the sales proportion of locally related products, are extracted from a list of locally related products for the 80 outlets. These two types of characteristics are then merged to form a classification feature set. Using data from 200 labeled fresh food commercial outlets in City A as training samples, a random forest outlet classification model is trained. The feature data of the 80 outlets to be classified are input into the trained model, and the model outputs classification results: 25 outlets are classified as specialized sales stores, and 55 outlets are classified as comprehensive retail outlets. The comprehensive retail outlets include 20 online fresh food e-commerce stores and 35 offline comprehensive supermarkets.

[0062] This step constructs a branch classification model by integrating the business characteristics of the branches with the characteristics of local product operations. This ensures that the classification results of the branches are deeply adapted to the statistical needs of local product transaction volume, effectively improving the accuracy and adaptability of the branch classification. The generated classification results can provide a reasonable premise for designing differentiated transaction volume calculation logic for different types of branches, making the transaction volume calculation more in line with the transaction characteristics of different types of branches.

[0063] Step 4: Based on the standardized dataset and the local associated product list, determine the local associated sales attributes of the professional sales stores, and calculate the local associated product transaction volume of the professional sales stores that meet the local associated sales attributes according to the determination results.

[0064] Among them, the local related sales attribute refers to the attribute of a professional sales store whose core business is to operate local related products.

[0065] Specialty stores are sales outlets for a single product category. Their local product transaction volume can be directly calculated based on transaction data. However, local small and micro specialty stores are prone to non-local fraudulent transactions, resulting in invalid data mixed in with the statistical results. In one possible implementation, step 4 includes the following sub-steps:

[0066] Step 4.1: Based on the standardized dataset and the local associated product list, determine the local associated sales attributes of the professional sales stores and filter out the professional sales stores that meet the local associated sales attributes.

[0067] In the specific implementation process, based on the professional sales store operation data in the standardized dataset and the product information in the local related product list, the professional sales store is judged to have local related sales attributes. If the proportion of local related product sales of a professional sales store reaches a preset ratio, the professional sales store is judged to have local related sales attributes. All professional sales stores that meet the local related sales attributes are screened out. The reference range for the preset ratio is 80%-95%, which can also be determined by the local industry authorities according to the characteristics of the industry.

[0068] Step 4.2: Extract the transaction entity location characteristics and logistics link location characteristics from the transaction behavior data corresponding to the professional sales stores that meet the location-related sales attributes.

[0069] Among them, the geographical characteristics of the transaction entity refer to the geographical characteristics of the buyer in the product transaction, including the buyer's registered address, the location of the payment mobile phone number, etc.; the geographical characteristics of the logistics link refer to the geographical characteristics of the logistics link in the product transaction, including the product's place of shipment, the place of receipt, etc.

[0070] In the specific implementation process, transaction behavior data corresponding to professional sales stores that conform to the local related sales attributes are extracted from the standardized dataset, and the local characteristics of the transaction subject and the logistics link of each transaction are extracted from the transaction behavior data.

[0071] Step 4.3: Verify the authenticity of the transaction behavior based on the location characteristics of the transaction entity and the location characteristics of the logistics link, and eliminate non-local fraudulent transaction data.

[0072] In the specific implementation process, the authenticity of the transaction behavior is verified based on the location characteristics of the transaction entity and the location characteristics of the logistics link. If the registered address and payment mobile phone number of the transaction entity are both outside the local area, and the logistics receiving place is outside the local area, and only the shipping place is within the local area, then the transaction is determined to be a non-local order-brushing transaction, and the non-local order-brushing transaction data generated by such transactions are removed.

[0073] Step 4.4: Based on the verified transaction behavior data, calculate the transaction amount of local related products of the professional sales stores that meet the local related sales attributes.

[0074] In the specific implementation process, the verified transaction behavior data after removing non-local order-brushing transaction data will be integrated, and the total amount of all local related product transactions of professional sales stores that meet the local related sales attributes will be counted, which is the local related product transaction amount of such professional sales stores.

[0075] For example, taking the fresh agricultural products sector in City A as an example, 25 specialized sales stores were assessed for their local related sales attributes. If the sales proportion of local related products was ≥90%, the stores were deemed to have this attribute. Ultimately, 20 specialized sales stores that met the local related sales attribute were selected. Transaction behavior data of these 20 outlets were extracted from a standardized dataset, and the location characteristics of the transaction entity and logistics link, such as the buyer's registered address and delivery location, were extracted for each transaction. The transaction data was verified, and bulk order data from purchasing agents whose buyer's registered address and delivery location were both outside the province were removed, resulting in the removal of 500 invalid transaction records. Based on the verified transaction behavior data, the transaction amount of local related products of the 20 specialized sales stores was calculated to be 8.6 million yuan.

[0076] This step involves determining the local sales attributes of professional stores, thereby screening those that primarily sell products from their local area. Simultaneously, it verifies the authenticity of transactions by checking the dual locality of the transaction entity and the logistics link, effectively identifying and eliminating non-local fraudulent transactions. The transaction volume statistics based on verified genuine transaction data effectively improve the accuracy of local sales volume statistics for professional stores and reduce the interference of invalid data on the statistical results.

[0077] Step 5: Based on the standardized dataset and the local associated product list, calculate the proportion of local associated product transaction volume of the comprehensive retail outlets using a weighted model, and calculate the local associated product transaction volume in combination with the overall transaction data of the outlets.

[0078] Considering the diverse product categories, mixed sales of local and out-of-town products, and cross-regional transactions in comprehensive retail outlets, the existing technology, which includes cross-regional transaction data in local transaction volume and uses a single indicator to calculate the proportion, is prone to overestimation and bias in the statistical results. In one possible implementation, step 5 includes the following sub-steps:

[0079] Step 5.1: Extract the location characteristics of the transaction entities from the transaction behavior data of comprehensive retail outlets based on the standardized dataset, and split the overall transaction data of the outlets into local transaction data and cross-regional transaction data according to the location characteristics of the transaction entities.

[0080] Local transaction data refers to transaction data of comprehensive retail outlets where the transaction subject is a local entity, reflecting the actual consumption situation within the local area; cross-regional transaction data refers to transaction data of comprehensive retail outlets where the transaction subject is a non-local entity.

[0081] In the specific implementation process, based on the comprehensive retail outlet transaction behavior data in the standardized dataset, the geographical location characteristics of the transaction entity for each transaction are extracted. The overall transaction data of the outlet is then split according to the geographical location characteristics of the transaction entity. If either the registered address or the mobile phone number of the transaction entity is located within the local area, the transaction data is classified as local transaction data. If neither the registered address nor the mobile phone number of the transaction entity is located outside the local area, the transaction data is classified as cross-regional transaction data.

[0082] Step 5.2: Based on the standardized dataset, the local associated product list, and the local transaction data, construct a weighted model to calculate the proportion of local transaction volume of the local associated products of the comprehensive retail outlets.

[0083] The weighted model refers to a mathematical model constructed based on multi-dimensional quantitative indicators of local related products, used to calculate the proportion of local transaction volume of local related products; the proportion of local transaction volume of local related products refers to the proportion of the transaction value of local related products to the total local transaction value of all products in comprehensive retail outlets.

[0084] In one possible implementation, constructing the weighted model includes the following:

[0085] Based on the local transaction data and the local associated product list, the raw data of four indicators—sales volume, product price, number of transactions, and number of valid comments—of both local and non-local associated products are extracted.

[0086] The raw data of the four indicators are normalized and outlier removal is performed to generate standardized indicator data for each product.

[0087] The entropy weight method is adopted, based on the standardized indicator data of local and non-local products. The objective weight is determined by calculating the information entropy of each indicator, and the weights of the four indicators are calculated. The four standardized indicator data of each product are multiplied by their corresponding weights and then summed to obtain the weighted value of each product.

[0088] The ratio of the weighted average value of all local related products to the weighted average value of all products in comprehensive retail outlets is used to obtain the local transaction volume share of local related products.

[0089] In the specific implementation process, based on local transaction data and local related product lists, the raw data of four indicators—sales quantity, product price, number of transactions, and number of valid reviews—of local related products and non-local products in comprehensive retail outlets are extracted. Outlier removal is performed on the raw data of the four indicators, removing indicator data that exceeds the reasonable range of the industry. Then, the Min-Max normalization method is used to process the raw data after removing outliers, mapping the indicator data to the range of 0 to 1, and generating standardized indicator data for each product.

[0090] The entropy weighting method is used to objectively assign weights to standardized indicator data. Let the number of product samples to be calculated be n, the number of indicators be m, and the standardized indicator data for the j-th indicator of the i-th product be... The entropy value of the j-th index The calculation formula is:

[0091] (1)

[0092] in , ,like ,but .

[0093] The weight of the j-th indicator The calculation formula is:

[0094] (2)

[0095] The formula is used to calculate the weights of four indicators: sales volume, product price, number of transactions, and number of valid reviews.

[0096] The formula for calculating the weighted value of a single product is as follows: (The formula is repeated in the original text.)

[0097] (3)

[0098] in This is the weighted value for a single product of the i-th type. , , , These are the four standardized indicator data for the i-th product.

[0099] Calculate the total weighted average value of all products related to the same location. The sum of the weighted average values ​​of all individual products in a comprehensive retail outlet The formula for calculating the proportion of local transaction volume for products with local territorial association is as follows:

[0100] (4)

[0101] Step 5.3: Combine the proportion of local transaction volume of the local related products with the local transaction data to calculate the transaction volume of local related products of the comprehensive retail outlet.

[0102] In the specific implementation process, the calculated proportion of local transaction volume P of local related products is multiplied by the total amount of local transaction data of comprehensive retail outlets, and the product is the transaction volume of local related products of comprehensive retail outlets.

[0103] For example, taking 55 comprehensive retail outlets in the fresh agricultural products sector of City A as an example, the location characteristics of the transaction entities are extracted from the transaction behavior data, and the overall transaction data of the outlets is divided into local transaction data and cross-regional transaction data, of which the total amount of local transaction data is 21.4 million yuan.

[0104] The raw data of four indicators—sales volume, product price, number of transactions, and number of valid reviews—of local and non-local products were extracted from 55 outlets. Data on imported fruits with abnormally high prices and fake reviews with a number of reviews far exceeding the number of transactions were removed. The remaining data underwent Min-Max normalization to generate standardized indicator data. The entropy weight method was used to calculate the weights of the four indicators as follows: number of transactions 0.4, sales volume 0.3, product price 0.2, and number of valid reviews 0.1. Based on the single-item weighting formula, the total weighted value of the 238 local products was calculated to be 820, and the total weighted value of all 320 products was 1930. The proportion of local transaction volume for local products was approximately 820 / 1930 ≈ 42.5%. Combining this proportion with local transaction data, the transaction volume of local products for the 55 comprehensive retail outlets was calculated to be 21.4 million × 42.5% = 9.195 million yuan.

[0105] This step effectively eliminates the interference of cross-regional transaction data on the calculation of local transaction volume by separating local and cross-regional transaction data of comprehensive retail outlets, thus avoiding the overestimation of statistical results. A weighted model is constructed based on four indicators: sales volume, product price, number of transactions, and number of valid reviews. The entropy weight method is used for objective weighting, which provides a clear theoretical basis and implementation process for calculating the proportion of local transaction volume of products related to the local area. This effectively improves the scientificity and accuracy of the proportion calculation, and the final calculated transaction volume of products related to the local area of ​​comprehensive retail outlets is more in line with the actual consumption situation in the local area.

[0106] Step 6: Summarize the transaction volume of local related products of the professional sales stores and comprehensive retail outlets, and conduct multi-dimensional statistical analysis according to the statistical dimensions of local industry benchmark data to generate local related product transaction volume statistics results.

[0107] Among them, multi-dimensional statistical analysis refers to the process of breaking down, statistically analyzing, and integrating the aggregated transaction volume of products related to the locality according to different statistical dimensions.

[0108] Considering that existing technologies only perform simple total statistics on local product transaction volume without standardized multi-dimensional statistical analysis, the reference value of the statistical results is limited and cannot meet the diverse data needs of regional industrial development and policy formulation, in one possible implementation, the multi-dimensional statistical analysis in step 6 includes the following sub-steps:

[0109] Step 6.1: Based on the statistical dimensions of the local industry benchmark data, break it down into industry dimension, region dimension, time dimension, and outlet type dimension.

[0110] Among them, the industry dimension refers to the dimension of calculating transaction volume according to the sub-sectors of the local industry; the geographical dimension refers to the dimension of calculating transaction volume according to the administrative division of the locality; the time dimension refers to the dimension of calculating transaction volume according to the time period; and the outlet type dimension refers to the dimension of calculating transaction volume according to the type of commercial outlet.

[0111] In the specific implementation process, statistical dimension classification standards are extracted from the local industry benchmark data. According to these standards, four core statistical dimensions are divided into four dimensions: industry dimension, region dimension, time dimension, and outlet type dimension. The industry dimension is divided according to the sub-sectors of the local industry, the region dimension is divided according to the administrative level of the local area, the time dimension is divided according to the preset time period, and the outlet type dimension is divided into professional sales stores, online comprehensive retail outlets, and offline comprehensive retail outlets.

[0112] Step 6.2: Decompose and statistically analyze the aggregated transaction amount of related products in each region according to each dimension to generate single-dimensional statistical data.

[0113] Among them, single-dimensional statistical data refers to the transaction amount data under each subdivision obtained after breaking down the aggregated transaction amount of products related to the locality into a single dimension.

[0114] In the specific implementation process, the total transaction amount of local related products after aggregation is broken down into industry, region, time and outlet type dimensions. The transaction amount under each subdivision is statistically analyzed one by one. For example, the transaction amount of each sub-industry under the industry dimension and the transaction amount of each administrative region under the region dimension are statistically analyzed. The statistical results are then compiled into single-dimensional statistical data.

[0115] Step 6.3: Perform cross-analysis on the statistical data of each single dimension to generate local related product transaction volume statistics.

[0116] Among them, the statistical results of local related product transaction volume refer to the multi-dimensional and multi-level statistical data set of local related product transaction volume generated after cross-integration analysis of single-dimensional statistical data.

[0117] In the specific implementation process, cross-integration analysis is carried out on the statistical data of each single dimension, and the subdivided data of different dimensions are combined to statistically analyze the transaction volume data under the combined dimensions. For example, the transaction volume of online comprehensive retail outlets in a certain administrative region and a certain sub-industry during a certain time period is statistically analyzed. All the cross-integration analysis results of the transaction volume data are integrated to generate the local related product transaction volume statistics results.

[0118] For example, taking the fresh agricultural products sector in City A as an example, the total transaction amount of 8.6 million yuan from 20 specialized sales stores and 9.195 million yuan from 55 comprehensive retail outlets related to local products is summarized to obtain a total of 17.795 million yuan. Based on the statistical dimensions of City A's industry benchmark data, the following dimensions are broken down: industry (strawberries, rice, handmade bean products, etc.), region (XX district, XX county, XX township in City A), time (daily and weekly for the past 30 days), and outlet type (specialized sales stores, online comprehensive retail outlets, offline comprehensive retail outlets). The total is then... The transaction amount of 50,000 yuan was broken down and statistically analyzed according to various dimensions to generate single-dimensional statistical data, such as 8.2 million yuan for the transaction amount of strawberry products in XX District, 6.5 million yuan for the transaction amount of strawberry products, and 8.6 million yuan for the transaction amount of specialty stores. Cross-integration analysis of the single-dimensional statistical data was conducted to obtain combined dimension data such as 850,000 yuan for the average weekly transaction amount of strawberry products in online comprehensive retail outlets in XX District in the past 30 days and 2.1 million yuan for the transaction amount of rice products in specialty stores in XX County in the past 30 days. All single-dimensional and combined dimension data were integrated to generate the statistical results of the transaction amount of local related products in the field of fresh agricultural products in City A.

[0119] This step, based on the statistical dimensions of local industry benchmark data, enables multi-dimensional breakdown, statistical analysis, and cross-integration of local related product transaction volume. The generated statistical results are clear in dimensions and rich in levels, reflecting not only the overall situation of local related product transaction volume but also accurately reflecting the transaction situation of various sub-sectors, administrative regions, and types of outlets. This effectively enhances the pertinence and practical application value of the statistical results, providing comprehensive and accurate data support for regional industrial development planning and the formulation of local product support policies.

[0120] Example 2

[0121] This embodiment provides a local product transaction volume calculation system based on multi-dimensional feature analysis. It applies the aforementioned local product transaction volume calculation method based on multi-dimensional feature analysis, such as... Figure 2 As shown, it includes:

[0122] The multi-source data acquisition and standardization module is used to collect basic data of commercial outlets, product attribute data, transaction behavior data, and local industry benchmark data, and to perform data standardization processing to generate a standardized dataset.

[0123] The product location feature analysis and determination module is used to extract the location association features of products based on the standardized dataset, match them with the local industry benchmark data to determine whether the products have location association attributes, and generate a local associated product list.

[0124] A business outlet intelligent classification module is used to classify business outlets based on the standardized dataset and the local associated product list, and generate classification results for specialty stores and general retail outlets;

[0125] The professional sales store location determination and transaction volume statistics module is used to determine the location-related sales attributes of the professional sales stores based on the standardized dataset and the local related product list, and to count the local related product transaction volume of professional sales stores that meet the location-related sales attributes according to the determination results.

[0126] A weighted calculation module for local transaction volume of comprehensive retail outlets is used to calculate the proportion of local related product transaction volume of comprehensive retail outlets based on the standardized dataset and the local related product list, and to calculate the local related product transaction volume in combination with the overall transaction data of the outlets.

[0127] The local product transaction volume multi-dimensional summary and statistics module is used to summarize the local related product transaction volume of the professional sales stores and comprehensive retail outlets, perform multi-dimensional statistical analysis according to the statistical dimensions of local industry benchmark data, and generate local related product transaction volume statistics results.

[0128] The embodiments described above are merely illustrative of specific implementations of the present invention, and while the descriptions are detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for calculating local product transaction volume based on multi-dimensional feature analysis, characterized in that, Includes the following steps: Step 1: Collect basic data of commercial outlets, product attribute data, transaction behavior data, and local industry benchmark data, and perform data standardization processing to generate a standardized dataset; Step 2: Extract the local association features of the products based on the standardized dataset, match them with the local industry benchmark data to determine whether the products have local association attributes, and generate a list of locally associated products. Step 3: Classify commercial outlets based on the standardized dataset and the local associated product list to generate classification results for specialty stores and general retail outlets; Step 4: Based on the standardized dataset and the local associated product list, determine the local associated sales attributes of the professional sales stores, and calculate the local associated product transaction amount of the professional sales stores that meet the local associated sales attributes according to the determination results. Step 5: Based on the standardized dataset and the local associated product list, calculate the proportion of local associated product transaction volume of the comprehensive retail outlets using a weighted model, and calculate the local associated product transaction volume in combination with the overall transaction data of the outlets; Step 6: Summarize the transaction volume of local related products of the professional sales stores and comprehensive retail outlets, and conduct multi-dimensional statistical analysis according to the statistical dimensions of local industry benchmark data to generate local related product transaction volume statistics results.

2. The method for calculating local product transaction volume based on multi-dimensional feature analysis according to claim 1, characterized in that, Step 2 includes the following sub-steps: Step 2.1: Extract the basic geographical association features and distinctive geographical association features of the product based on the standardized dataset. The basic geographical association features include trademarks, brands, and production location codes. The distinctive geographical association features include geographical industry identifiers, product distribution links, and production process attribute features. Step 2.2: Perform the initial matching of the basic geographical location association features with the geographical location industry benchmark data; Step 2.3: For products that are not determined to have geographical association attributes in the first match, perform a second match between the distinctive geographical association features and the geographical industry benchmark data; Step 2.4: Based on the results of the first and second matching, determine whether the product has a local association attribute and generate a list of locally associated products.

3. The method for calculating local product transaction volume based on multi-dimensional feature analysis according to claim 1, characterized in that, Step 3 includes the following sub-steps: Step 3.1: Extract the business characteristics of the business outlets based on the standardized dataset, and extract the local product business characteristics of the business outlets based on the local related product list; Step 3.2: Construct a network classification model by integrating the self-operating characteristics with the local product operating characteristics; Step 3.3: Input the feature data of the commercial outlets into the outlet classification model to complete the classification of specialty stores and general retail outlets, and generate the classification results.

4. The method for calculating local product transaction volume based on multi-dimensional feature analysis according to claim 1, characterized in that, Step 4 includes the following sub-steps: Step 4.1: Based on the standardized dataset and the local associated product list, determine the local associated sales attributes of the professional sales stores and filter out the professional sales stores that meet the local associated sales attributes. Step 4.2: Extract the location characteristics of the transaction entities and the location characteristics of the logistics links from the transaction behavior data corresponding to the professional sales stores that conform to the location-related sales attributes; Step 4.3: Verify the authenticity of the transaction behavior based on the location characteristics of the transaction entity and the location characteristics of the logistics link, and eliminate non-local fraudulent transaction data; Step 4.4: Based on the verified transaction behavior data, calculate the transaction amount of local related products of the professional sales stores that meet the local related sales attributes.

5. The method for calculating local product transaction volume based on multi-dimensional feature analysis according to claim 1, characterized in that, Step 5 includes the following sub-steps: Step 5.1: Extract the location characteristics of the transaction entities from the transaction behavior data of comprehensive retail outlets based on the standardized dataset, and split the overall transaction data of the outlets into local transaction data and cross-regional transaction data according to the location characteristics of the transaction entities; Step 5.2: Based on the standardized dataset, the local associated product list, and the local transaction data, construct a weighted model to calculate the proportion of local transaction volume of the local associated products of the comprehensive retail outlets; Step 5.3: Combine the proportion of local transaction volume of the local related products with the local transaction data to calculate the transaction volume of local related products of the comprehensive retail outlet.

6. The method for calculating local product transaction volume based on multi-dimensional feature analysis according to claim 1 or 5, characterized in that, The construction of the weighted model includes the following: Based on the local transaction data and the local associated product list, the raw data of four indicators—sales quantity, product price, number of transactions, and number of valid comments—of local associated products and non-local products are extracted. The raw data of the four indicators are normalized and outlier removal is performed to generate standardized indicator data for each product. The standardized indicator data are objectively weighted using the entropy weight method to calculate the weights of the four indicators; the four standardized indicator data of each product are multiplied by their corresponding weights and then summed to obtain the weighted value of each product. The ratio of the weighted average value of all local related products to the weighted average value of all products in comprehensive retail outlets is used to obtain the local transaction volume share of local related products.

7. The method for calculating local product transaction volume based on multi-dimensional feature analysis according to claim 1, characterized in that, Step 6, the multidimensional statistical analysis, includes the following sub-steps: Step 6.1: Based on the statistical dimensions of the local industry benchmark data, break it down into industry dimension, region dimension, time dimension, and outlet type dimension; Step 6.2: Decompose and statistically analyze the aggregated transaction amount of related products in the region according to each dimension to generate single-dimensional statistical data; Step 6.3: Perform cross-analysis on the statistical data of each single dimension to generate local related product transaction volume statistics.

8. A local product transaction volume calculation system based on multi-dimensional feature analysis, characterized in that, The method for calculating local product transaction volume based on multi-dimensional feature analysis as described in any one of claims 1-7 includes: The multi-source data acquisition and standardization module is used to collect basic data of commercial outlets, product attribute data, transaction behavior data, and local industry benchmark data, and to perform data standardization processing to generate a standardized dataset. The product location feature analysis and determination module is used to extract the location association features of products based on the standardized dataset, match them with the local industry benchmark data to determine whether the products have location association attributes, and generate a local associated product list. A business outlet intelligent classification module is used to classify business outlets based on the standardized dataset and the local associated product list, and generate classification results for specialty stores and general retail outlets; The professional sales store location determination and transaction volume statistics module is used to determine the location-related sales attributes of the professional sales stores based on the standardized dataset and the local related product list, and to count the local related product transaction volume of professional sales stores that meet the location-related sales attributes according to the determination results. A weighted calculation module for local transaction volume of comprehensive retail outlets is used to calculate the proportion of local related product transaction volume of comprehensive retail outlets based on the standardized dataset and the local related product list, and to calculate the local related product transaction volume in combination with the overall transaction data of the outlets. The local product transaction volume multi-dimensional summary and statistics module is used to summarize the local related product transaction volume of the professional sales stores and comprehensive retail outlets, perform multi-dimensional statistical analysis according to the statistical dimensions of local industry benchmark data, and generate local related product transaction volume statistics results.

Images