Risk rating prediction system based on e-commerce big data analysis

Abstract

The invention discloses a risk rating prediction system based on e-commerce big data analysis, and relates to the technical field of e-commerce. A first data acquisition module discloses an API interface through an e-commerce platform, and collects exposure frequency Pbg and selling quantity data of a target commodity in each time period; the correlation analysis module generates an exposure correlation index Zsm based on the correlation between the target commodity exposure frequency data and the target commodity selling quantity data so as to preliminarily judge whether the target commodity has selling competitiveness or not, and sends a selling risk analysis instruction; the second data acquisition module acquires target commodity selling price information data and competitive shop after-sales information data according to the selling risk analysis instruction, the comprehensive selling risk assessment module performs fitting to obtain a comprehensive selling risk assessment index Zpg by using a machine deep learning technology, the risk rating prediction module compares the comprehensive selling risk assessment index Zpg with a preset threshold P, and the risk rating prediction module performs risk evaluation. And predicting a corresponding selling risk level, and making corresponding push information.

Description

Technical Field

[0001] The present invention relates to the technical field of e-commerce, and in particular to a risk rating prediction system based on e-commerce big data analysis. Background Art

[0002] With the rapid development of the e-commerce industry, online sales of goods have become the main profit channel for major companies. In this process, how to effectively manage and optimize the sales strategy of goods has become an important issue for e-commerce platforms and merchants. Multiple factors such as product exposure frequency, sales volume, price fluctuations and after-sales service directly affect the market competitiveness and profitability of goods. However, the complexity and variability of the market environment make merchants face many risks when selling goods, including changes in product exposure traffic, fluctuations in consumer demand and price sensitivity. These risks not only affect the sales effect of goods, but may also lead to inventory backlogs and waste of resources. Therefore, it is particularly important to predict the risk level of product sales.

[0003] At present, most e-commerce companies rely on historical sales data and subjective judgment when predicting the risk level of commodity sales, which lacks systematicity and scientificity. This traditional method is not only inefficient, but also easily affected by human factors, resulting in inaccurate risk level prediction. In addition, the lack of real-time data collection and analysis mechanism makes it difficult for merchants to respond to market changes in a timely manner, thus missing the best sales opportunities. Therefore, a risk rating prediction system based on e-commerce big data analysis is constructed. Through multi-level data collection and analysis, it can comprehensively consider exposure frequency, sales quantity, price fluctuations and after-sales information factors, provide e-commerce companies and merchants with more accurate commodity sales risk rating prediction solutions, help them identify potential problems in a timely manner and take corresponding measures, thereby effectively reducing sales risks and improving market competitiveness. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a risk rating prediction system based on e-commerce big data analysis, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a risk rating prediction system based on e-commerce big data analysis, comprising a first data collection module, a correlation analysis module, a second data collection module, a comprehensive sales risk assessment module and a risk rating prediction module; The first data collection module is used to collect the exposure frequency data of the target product in each set time period by accessing the public API interface of the e-commerce platform to obtain the exposure frequency Pbg of the target product, and collect the sales quantity data of the target product in the corresponding time period according to the sales quantity of the target product; The correlation analysis module is used to analyze the impact of the target product exposure frequency Pbg on the target product sales quantity Ssm according to the correlation between the target product exposure frequency data and the target product sales quantity data, generate the exposure correlation index Zsm, preliminarily determine whether the target product has sales competitiveness on the e-commerce platform, and issue corresponding sales risk analysis instructions; The second data collection module is used to collect the target commodity selling price information data of each period on the e-commerce platform after receiving the sales risk analysis instruction, and collect the target commodity after-sales information according to the operating conditions of the stores selling the same target commodity on the e-commerce platform, and obtain the after-sales information data of the competing stores; The comprehensive sales risk assessment module is used to obtain the comprehensive sales risk assessment index Zpg by fitting based on the target product exposure frequency data, target product sales quantity data, target product sales price information and competing store after-sales information data by using machine deep learning technology, and performing feature extraction and dimensionless processing; The risk rating prediction module is used to compare and analyze the preset evaluation threshold value P with the comprehensive sales risk evaluation index Zfx, predict the corresponding sales risk level, and make corresponding push information according to the sales risk level.

[0006] Preferably, the first data acquisition module includes an access unit and a data acquisition unit; The access unit is used to register an e-commerce platform developer account, obtain an API key and document, identify available API endpoints and request formats, and use the programming language Python to construct HTTP request related parameters through a library, wherein the related parameters include a target product ID and a search keyword; The data collection unit is based on the public API interface of the connected e-commerce platform, sends a search request, receives response data, and collects the target product exposure frequency data and the target product sales quantity data in each time period after feature recognition and screening, wherein the target product exposure frequency data includes the target product exposure frequency Pbg in each time period, and the target product sales quantity data includes the target product sales quantity Ssm in each time period.

[0007] Preferably, the correlation analysis module includes an exposure correlation analysis unit and a judgment unit; The exposure correlation analysis unit is used to analyze the influence of the target commodity exposure frequency Pbg on the target commodity sales quantity Ssm based on the machine learning model and in combination with the correlation between the target commodity exposure frequency data and the target commodity sales quantity data, so as to generate an exposure correlation index Zsm. The exposure correlation index Zsm is obtained by the following formula: ; In the formula, It is represented as the exposure frequency of the target product in the i-th collection period, It is represented by the sales quantity of the target product in the i-th collection period. It is expressed as the average exposure frequency of the target product during the collection period. It is represented as the mean sales quantity of the target product during the collection period, i=1, 2, 3, ..., n, where n is the collection period.

[0008] Preferably, the judgment unit is used to preliminarily judge whether the target product has sales competitiveness on the e-commerce platform according to the value corresponding to the exposure correlation index Zsm, and issue a corresponding sales risk analysis instruction, the specific contents of which are as follows: If the exposure correlation index Zsm=1, it means that there is a completely positive correlation between the exposure frequency and the sales quantity of the target product. It is preliminarily judged that the target product has sales competitiveness on the e-commerce platform, and a sales risk analysis instruction is issued; If the exposure correlation index Zsm=-1, it means that there is a completely negative correlation between the exposure frequency and the sales quantity of the target product. It is preliminarily judged that the target product does not have sales competitiveness on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit; If the exposure correlation index Zsm=0, it means that there is no correlation between the exposure frequency and sales quantity of the target product. It is preliminarily judged that the target product does not have sales competitiveness on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit.

[0009] Preferably, the second data collection module is used to collect target commodity selling price information data for each time period on the e-commerce platform after receiving the sales risk analysis instruction, the target commodity selling price information data including the target commodity selling price Gsm for the corresponding time period, and collect after-sales information of the target commodity based on the operating conditions of the stores selling the same target commodity on the e-commerce platform, and obtain after-sales information data of competing stores, the competing store after-sales information data including the target commodity return rate Vth and the target commodity negative review rate Vcp.

[0010] Preferably, the comprehensive sales risk assessment module includes an exposure conversion analysis unit, a price fluctuation analysis unit, a post-sales analysis unit and a sales risk assessment unit; The exposure conversion analysis unit is based on the exposure frequency data of the target product and the sales quantity data of the target product. When the exposure correlation index Zsm=1, there is a completely positive correlation between the exposure frequency and the sales quantity of the target product. After dimensionless processing, the conversion effect between the exposure frequency and the sales quantity of the target product in each time period is analyzed to obtain the exposure conversion coefficient Xzh. The exposure conversion coefficient Xzh is obtained by the following formula; ; In the formula, It is represented as the exposure frequency of the target product in the i-th collection period, It is represented by the sales quantity of the target product in the i-th collection period, i=1, 2, 3, ..., n, and n represents the collection cycle.

[0011] Preferably, the price fluctuation analysis unit is used to summarize the target commodity selling price information data to construct a target commodity selling price data set, and extract the highest target commodity selling price Gsm Max , through the statistical mean algorithm, obtain the average target product selling price in the set time period , and after linear normalization, the price fluctuation coefficient Xbd is obtained by fitting. The price fluctuation coefficient Xbd is obtained by the following formula: ; In the formula, i=1, 2, 3, ..., n, n represents the acquisition period, It is represented as the highest selling price of the target product in the i-th collection time period.

[0012] Preferably, the after-sales analysis unit is used to analyze the after-sales information data of the competing stores, and obtain the after-sales risk coefficient Xsh after dimensionless processing. The after-sales risk coefficient Xsh is obtained by the following formula: ; In the formula, in the formula, It is represented as the target product return rate of the j-th store, It is represented as the negative review rate of the target product in the jth store, j=1, 2, 3, ..., m, where m is the number of stores. and are all expressed as weight values.

[0013] Preferably, the sales risk assessment unit is used to utilize machine deep learning technology, and dimensionlessly process the exposure conversion coefficient Xzh, the price fluctuation coefficient Xbd, and the after-sales risk coefficient Xsh, and obtain a comprehensive sales risk assessment index Zpg by fitting. The comprehensive sales risk assessment index Zpg is obtained by the following formula: ; In the formula, , and They are respectively represented as the weight values ​​of exposure conversion coefficient Xzh, price fluctuation coefficient Xbd and after-sales risk coefficient Xsh, and A is represented as the correction constant.

[0014] Preferably, the risk rating prediction module includes a rating prediction unit and a push unit; The rating prediction unit is used to compare and analyze the preset evaluation threshold value P with the comprehensive sales risk evaluation index Zpg, predict the corresponding sales risk level, and make corresponding push prompts according to the sales risk level. The specific contents are as follows; If the comprehensive sales risk assessment index Zpg ≥ the assessment threshold P, it means that the target product has sales competitiveness on the e-commerce platform, but the target product is in a risky sales state on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit; If the comprehensive sales risk assessment index Zpg is less than the assessment threshold P, it indicates that the target product has sales competitiveness on the e-commerce platform, and the target product is not in a risky sales state on the e-commerce platform, and the secondary sales risk level is obtained and sent to the push unit; The push unit is used to receive the primary selling risk level and the secondary selling risk level, and send the corresponding push information to the display screen, the specific content is as follows; When receiving the first-level sales risk level, the push information content is that the current target product is in a risky sales state. It is recommended that the merchant first implement competitive pricing and dynamic pricing strategies for the target product, appropriately adjust the price of the target product to be lower than the average price of the platform, ensure timely response to customer inquiries, and solve post-purchase problems; at the same time, pay attention to consumers' evaluation and feedback on the target product, take corrective measures to improve product quality and service for negative reviews, and then sell the target product on the e-commerce platform; When receiving the second-level sales risk level, the pushed information content is that the current target product is competitive in sales on the e-commerce platform and is not in a risky sales state. It is recommended that merchants sell the target product on the e-commerce platform. At the same time, the information on the product page is updated and optimized, including high-quality pictures, product descriptions and user reviews. According to user search habits, the product title and keywords are described in detail to increase natural search traffic, and marketing promotion is carried out through multiple online channels to attract potential customers to pay attention to the target product.

[0015] The present invention provides a risk rating prediction system based on e-commerce big data analysis, which has the following beneficial effects: (1) Through the collaborative work of multiple modules, an accurate assessment and prediction of the sales risk of the target product is achieved. First, the first data collection module obtains the exposure frequency and sales quantity of the target product in a timely manner through the API interface of the e-commerce platform to ensure the real-time and accuracy of the data. Then, based on the correlation between the exposure frequency data of the target product and the sales quantity data of the target product, the impact of the exposure frequency of the target product on the sales quantity is deeply analyzed, and the exposure correlation index Zsm is generated to preliminarily judge the market competitiveness of the product and issue risk analysis instructions. This process enables merchants to quickly identify potential sales problems of the target product. If there is no completely positive correlation between the exposure frequency and sales quantity of the target product, it means that the target product has no market competitiveness on the e-commerce platform. The second data collection module further collects the target product sales price Gsm and the after-sales information of competing stores. To provide more comprehensive data support for subsequent comprehensive sales risk assessment, the comprehensive sales risk assessment module integrates various types of data by applying machine deep learning technology, including exposure frequency, sales quantity, price fluctuations and after-sales service factors, to generate a comprehensive sales risk assessment index Zpg, making the prediction results more scientific and reliable. Finally, the risk rating prediction module compares the preset assessment threshold P with the comprehensive sales risk assessment index Zpg to accurately predict the sales risk level of the product and promptly push the corresponding adjustment strategy to the merchant. The implementation of this system not only improves the decision-making efficiency and market adaptability of the merchants, but also effectively reduces sales risks, improves market competitiveness, and reduces economic losses caused by blind sales. Ultimately, it provides merchants with a more accurate product sales risk rating prediction solution to ensure their steady development in the highly competitive e-commerce environment.

[0016] (2) Through the close collaboration between the exposure correlation analysis unit and the judgment unit, the market competitiveness assessment capability of the target products on the e-commerce platform is effectively improved. By using advanced machine learning models, the correlation between the exposure frequency and sales volume of the target products is analyzed to generate the exposure correlation index Zsm, which provides a quantitative basis for decision-making. When the exposure correlation index Zsm is 1, the system clearly shows the strong competitiveness of the product in the market, and merchants can quickly adjust their marketing strategies to increase sales. On the contrary, if the exposure correlation index Zsm is -1 or 0, the system will promptly identify potential risk products and give merchants a first-level sales risk level prompt to help them avoid possible economic losses. This dynamic risk analysis and timely feedback mechanism not only improves the merchants' response speed and decision-making quality, but also effectively reduces the risk of blind sales caused by market changes, thereby enabling merchants to more calmly respond to challenges in the increasingly competitive e-commerce environment and ensure the stability and growth of profitability.

[0017] (3) By accurately collecting the selling price information of the target products and the after-sales data of competing stores, the system can comprehensively evaluate market dynamics and ensure that merchants have a clear understanding of the product prices and competition. At the same time, the combination of exposure conversion analysis, price fluctuation analysis and after-sales analysis enables the system to deeply explore the conversion effect between product exposure and sales, the impact of price fluctuations and the after-sales performance of competitors. This comprehensive risk assessment mechanism not only improves the merchants' sensitivity to market changes, but also provides merchants with a scientific decision-making basis through the comprehensive sales risk assessment index Zpg generated by machine deep learning technology. The risk rating prediction module ensures that merchants can understand the market risk status of the products in a timely manner and adjust their sales strategies according to the corresponding push information, thereby maximizing profits and reducing potential losses. This set of systematic evaluation, prediction and feedback mechanisms not only improves the market competitiveness of merchants, but also provides them with the ability to respond flexibly in a complex e-commerce environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a block diagram of a risk rating prediction system based on e-commerce big data analysis in the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Example 1 See also Figure 1 , the present invention provides a risk rating prediction system based on e-commerce big data analysis, including a first data collection module, a correlation analysis module, a second data collection module, a comprehensive sales risk assessment module and a risk rating prediction module; The first data collection module is used to collect the exposure frequency data of the target product in each set time period by accessing the public API interface of the e-commerce platform to obtain the exposure frequency Pbg of the target product, and collect the sales quantity data of the target product in the corresponding time period according to the sales quantity of the target product; The correlation analysis module is used to analyze the impact of the target product exposure frequency Pbg on the target product sales quantity Ssm according to the correlation between the target product exposure frequency data and the target product sales quantity data, generate the exposure correlation index Zsm, preliminarily determine whether the target product has sales competitiveness on the e-commerce platform, and issue corresponding sales risk analysis instructions; The second data collection module is used to collect the target commodity selling price information data of each period on the e-commerce platform after receiving the sales risk analysis instruction, and collect the target commodity after-sales information according to the operating conditions of the stores selling the same target commodity on the e-commerce platform, and obtain the after-sales information data of the competing stores; The comprehensive sales risk assessment module is used to obtain the comprehensive sales risk assessment index Zpg by fitting based on the target product exposure frequency data, target product sales quantity data, target product sales price information and competing store after-sales information data by using machine deep learning technology, and performing feature extraction and dimensionless processing; The risk rating prediction module is used to compare and analyze the preset evaluation threshold value P with the comprehensive sales risk evaluation index Zfx, predict the corresponding sales risk level, and make corresponding push information according to the sales risk level.

[0021] In this embodiment, a comprehensive prediction and evaluation of the sales risk of target commodities on the e-commerce platform is provided through multi-layer data collection and analysis modules. First, the system integrates multiple types of data collection modules, including exposure frequency, sales quantity, price fluctuations and after-sales feedback key data, and effectively constructs a panoramic view of the commodity on the e-commerce platform. Secondly, based on the machine learning model and combining the correlation between the target commodity exposure frequency data and the target commodity sales quantity data, the system analyzes the impact of the target commodity exposure frequency Pbg on the target commodity sales quantity, and generates an exposure correlation index Zsm, so as to preliminarily judge the market competitiveness of the commodity and issue a risk analysis instruction. This process enables merchants to quickly identify potential sales problems of target commodities. If there is no completely positive correlation between the exposure frequency and sales quantity of the target commodity, it means that the target commodity has no market on the e-commerce platform. Competitiveness helps companies focus their promotional resources on high-potential products. In addition, the comprehensive sales risk assessment module generates a comprehensive sales risk assessment index Zpg representing the target product by extracting and processing features from multi-dimensional data. The comprehensive sales risk assessment index Zpg can comprehensively reflect the potential risks of selling products on e-commerce platforms. This function improves the efficiency and accuracy of risk rating predictions. Finally, the risk rating prediction module can quickly identify high-risk and low-risk products by comparing them with the pre-set assessment threshold P, and generate push information based on the results to help companies optimize decisions and respond to potential risks in advance. Through the above analysis process, companies can not only manage and adjust product sales strategies more scientifically and reduce operating risks, but also improve market competitiveness and resource utilization efficiency, ensure the safe sale of target products on e-commerce platforms, and improve benefits.

[0022] Example 2 Please refer to Figure 1, specifically: the first data acquisition module includes an access unit and a data acquisition unit; The access unit is used to register an e-commerce platform developer account, obtain an API key and document, identify available API endpoints and request formats, and use the programming language Python to construct HTTP request related parameters through a library, wherein the related parameters include a target product ID and a search keyword; It should be noted that the search keywords include, but are not limited to, keyword search content in terms of exposure data, sales quantity, sales price, return quantity and after-sales evaluation of the target product; The data collection unit is based on the public API interface of the connected e-commerce platform, sends a search request, receives response data, and collects the target product exposure frequency data and the target product sales quantity data in each time period after feature recognition and screening, wherein the target product exposure frequency data includes the target product exposure frequency Pbg in each time period, and the target product sales quantity data includes the target product sales quantity Ssm in each time period.

[0023] The target product exposure frequency Pbg refers to the number and frequency of times the target product is seen by users on the e-commerce platform during the corresponding period, which reflects the visibility and attractiveness of the target product on the platform, and is very important for analyzing its sales potential and competitiveness. The higher the exposure frequency, the more display opportunities the target product has on the platform, thereby attracting more potential consumers; the target product sales quantity Ssm refers to the total sales quantity of the target product in all stores on the e-commerce platform during the corresponding period.

[0024] In this embodiment, the efficiency and accuracy of data acquisition on the e-commerce platform are improved through the effective collaboration of the access unit and the data collection unit. The design of the access unit enables developers to easily register and obtain API keys, thereby quickly identifying available API endpoints and request formats, ensuring the smoothness of the data collection process. In addition, the use of Python programming language and related libraries to construct HTTP requests makes the data collection process highly flexible and adaptable. By sending search requests, the data collection unit can quickly receive and process response data. After feature recognition and screening, the system can accurately obtain the target product exposure frequency data Pbg and sales quantity data Ssm in each time period. This process not only improves the real-time nature of the data, but also ensures the quality and reliability of the data. More importantly, by collecting the exposure frequency and sales quantity of the target product, merchants can deeply analyze the market performance of the product and its potential sales opportunities, thereby providing strong data support for subsequent decision-making.

[0025] Example 3 Please refer to Figure 1,Specifically: the correlation analysis module includes an exposure correlation analysis unit and a judgment unit; The exposure correlation analysis unit is used to analyze the influence of the target commodity exposure frequency Pbg on the target commodity sales quantity Ssm based on the machine learning model and in combination with the correlation between the target commodity exposure frequency data and the target commodity sales quantity data, so as to generate an exposure correlation index Zsm. The exposure correlation index Zsm is obtained by the following formula: ; In the formula, It is represented as the exposure frequency of the target product in the i-th collection period, It is represented by the sales quantity of the target product in the i-th collection period. It is expressed as the average exposure frequency of the target product during the collection period. It is represented as the mean sales quantity of the target product during the collection period, i=1, 2, 3, ..., n, where n is the collection period.

[0026] Specifically, the judgment unit is used to preliminarily judge whether the target product has sales competitiveness on the e-commerce platform according to the value corresponding to the exposure correlation index Zsm, and issue a corresponding sales risk analysis instruction, the specific contents of which are as follows: If the exposure correlation index Zsm=1, it means that there is a completely positive correlation between the exposure frequency and the sales quantity of the target product, that is, the change direction of the two is completely consistent. It is preliminarily judged that the target product has sales competitiveness on the e-commerce platform, and a sales risk analysis instruction is issued; If the exposure correlation index Zsm=-1, it means that there is a completely negative correlation between the exposure frequency and the sales quantity of the target product, that is, the change directions of the two are completely opposite. It is preliminarily judged that the target product does not have sales competitiveness on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit; If the exposure correlation index Zsm=0, it means that there is no correlation between the exposure frequency and the sales quantity of the target product, that is, there is no linear relationship between the changes of the two. It is preliminarily judged that the target product does not have sales competitiveness on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit.

[0027] In this embodiment, through the efficient combination of the exposure correlation analysis unit and the judgment unit, support is provided for the product competitiveness evaluation of the e-commerce platform. The exposure correlation analysis unit uses a machine learning model to deeply explore the complex relationship between the exposure frequency Pbg of the target product and the sales quantity Ssm of the target product, and generates an exposure correlation index Zsm, thereby realizing accurate analysis of market dynamics. By comparing and analyzing data in different collection periods, not only can the relationship between the exposure frequency Pbg of the target product and the sales quantity Ssm of the target product be quantified, but also the data processing process can be simplified through a formulaic method, thereby improving the efficiency and accuracy of data analysis. Furthermore, according to the value range (-1, 1) of the exposure correlation index Zsm, the target product exposure frequency Pbg and the sales quantity Ssm of the target product can be intuitively reflected. The correlation between the product exposure frequency Pbg and the target product sales quantity Ssm is used to quickly judge the competitiveness of the target product on the e-commerce platform. If the exposure correlation index Zsm=1, it means that the exposure frequency and sales quantity of the target product are completely positively correlated, which means that the market demand is good and the target product is competitive on the e-commerce platform, and relevant risk analysis instructions are issued in time; on the contrary, if the exposure correlation index Zsm=-1 and the exposure correlation index Zsm=0, potential sales risks can be quickly identified to help merchants adjust their strategies in time to avoid losses. The implementation of this process can improve the response speed of e-commerce companies to market changes, ensure that they seize business opportunities and reduce risks in a highly competitive environment, thereby achieving more accurate market positioning and decision-making support.

[0028] Example 4 Please refer to Figure 1 Specifically: the second data collection module is used to collect target commodity selling price information data in each time period on the e-commerce platform after receiving the sales risk analysis instruction, the target commodity selling price information data includes the target commodity selling price Gsm of the corresponding time period, and collects after-sales information of the target commodity according to the operating conditions of the stores selling the same target commodity on the e-commerce platform, and obtains after-sales information data of competing stores, the competing stores after-sales information data includes the target commodity return rate Vth and the target commodity negative review rate Vcp.

[0029] In this embodiment, after receiving the sales risk analysis instruction, the sales price information and related after-sales data of the target product are quickly extracted from the e-commerce platform. The implementation of this module can not only obtain the target product sales price Gsm in each time period in real time, but also analyze the after-sales information of competing stores in combination with the operating conditions of similar target products in the market, including the target product return rate Vth and the target product negative review rate Vcp. This data collection method allows merchants to have an in-depth understanding of the market performance and consumer feedback of the target products, and provides a scientific basis for decision-making. In particular, the monitoring of high return rates and high negative review rates can help merchants discover problems in product quality and service links, thereby optimizing products and improving customer satisfaction. In addition, real-time price monitoring helps merchants adjust their pricing strategies to better cope with market competition. In summary, the second data acquisition module provides merchants with important market insights and decision-making support through accurate data acquisition and analysis, thereby reducing operating risks.

[0030] Example 5 Please refer to Figure 1 ,Specifically: the comprehensive sales risk assessment module includes an exposure conversion analysis unit, a price fluctuation analysis unit, a post-sales analysis unit and a sales risk assessment unit; The exposure conversion analysis unit is based on the exposure frequency data of the target product and the sales quantity data of the target product. When the exposure correlation index Zsm=1, there is a completely positive correlation between the exposure frequency and the sales quantity of the target product. After dimensionless processing, the conversion effect between the exposure frequency and the sales quantity of the target product in each time period is analyzed to obtain the exposure conversion coefficient Xzh. The exposure conversion coefficient Xzh is obtained by the following formula; ; In the formula, It is represented as the exposure frequency of the target product in the i-th collection period, It is represented by the sales quantity of the target product in the i-th collection period, i=1, 2, 3, ..., n, and n represents the collection cycle.

[0031] Specifically, the price fluctuation analysis unit is used to construct a target commodity selling price data set based on the target commodity selling price information data, and extract the highest target commodity selling price Gsm Max , through the statistical mean algorithm, obtain the average target product selling price in the set time period , and after linear normalization, the price fluctuation coefficient Xbd is obtained by fitting. The price fluctuation coefficient Xbd is obtained by the following formula: ; In the formula, i=1, 2, 3, ..., n, n represents the acquisition period, It is represented as the highest selling price of the target product in the i-th collection time period.

[0032] Specifically, the after-sales analysis unit is used to analyze the after-sales information data of the competing store, and obtain the after-sales risk coefficient Xsh after dimensionless processing. The after-sales risk coefficient Xsh is obtained by the following formula: ; In the formula, in the formula, It is represented as the target product return rate of the j-th store, It is represented as the negative review rate of the target product in the jth store, j=1, 2, 3, ..., m, where m is the number of stores. and are all expressed as weight values.

[0033] In this embodiment, through the collaborative work of each unit, in-depth and comprehensive market insights are provided, thereby helping merchants to formulate scientific sales strategies in the fiercely competitive e-commerce environment. First, the exposure conversion analysis unit uses the conversion relationship between the target product exposure frequency Pbg and the target product sales quantity Ssm to evaluate the conversion effect of the product in the market and calculate the exposure conversion coefficient Xzh, providing merchants with direct feedback on advertising effects and actual sales. This analysis can guide subsequent marketing decisions. Secondly, the price fluctuation analysis unit extracts the highest target product selling price Gsm by summarizing and counting the selling price data of the target product. Max And calculate the average target product selling price , and further obtain the price fluctuation coefficient Xbd. This data analysis helps merchants understand the dynamic changes of market prices and optimize pricing strategies to improve competitiveness. Finally, the after-sales analysis unit calculates the after-sales risk coefficient Xsh by evaluating the target product return rate Vth and the target product bad review rate Vcp of competing stores, providing merchants with an in-depth understanding of product quality and customer satisfaction. Efficient after-sales management can not only reduce the return rate, but also improve consumer experience, thereby enhancing the brand image. In summary, the above analysis process, through the combination of exposure conversion analysis, price fluctuation analysis and after-sales analysis, enables the system to deeply explore the conversion effect between product exposure and sales, the impact of price fluctuations and the after-sales performance of competitors. This comprehensive risk assessment mechanism not only improves merchants' sensitivity to market changes, but also provides a basis for scientific decision-making.

[0034] Example 6 Please refer to Figure 1Specifically: the sales risk assessment unit is used to utilize machine deep learning technology, and dimensionlessly process the exposure conversion coefficient Xzh, the price fluctuation coefficient Xbd, and the after-sales risk coefficient Xsh, and obtain a comprehensive sales risk assessment index Zpg by fitting. The comprehensive sales risk assessment index Zpg is obtained by the following formula; ; In the formula, , and They are respectively expressed as the weight values ​​of exposure conversion coefficient Xzh, price fluctuation coefficient Xbd and after-sales risk coefficient Xsh, 0≤ ≤1,0≤ ≤1,0≤ ≤1, and + + =1, A represents the correction constant, and the weight value can be obtained by referring to the hierarchical analysis method.

[0035] It should be noted that the analytic hierarchy process is a decision-making method that combines qualitative and quantitative analysis. It breaks down complex problems into multiple levels and determines their relative weights by comparing the importance of factors at each level. By comparing and calculating the weights layer by layer, it helps decision makers conduct in-depth analysis of complex problems and ultimately determine the best decision-making plan. In e-commerce data analysis, the analytic hierarchy process can be used to determine the weight value of each indicator, thereby more accurately evaluating the comprehensive sales risk of goods.

[0036] Specifically, the risk rating prediction module includes a rating prediction unit and a push unit; The rating prediction unit is used to compare and analyze the preset evaluation threshold value P with the comprehensive sales risk evaluation index Zpg, predict the corresponding sales risk level, and make corresponding push prompts according to the sales risk level. The specific contents are as follows; If the comprehensive sales risk assessment index Zpg ≥ the assessment threshold P, it means that the target product has sales competitiveness on the e-commerce platform, but the target product is in a risky sales state on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit; If the comprehensive sales risk assessment index Zpg is less than the assessment threshold P, it indicates that the target product has sales competitiveness on the e-commerce platform, and the target product is not in a risky sales state on the e-commerce platform, and the secondary sales risk level is obtained and sent to the push unit; The push unit is used to receive the primary selling risk level and the secondary selling risk level, and send the corresponding push information to the display screen, the specific content is as follows; When receiving the first-level sales risk level, the push information content is that the current target product is in a risky sales state. It is recommended that the merchant first implement competitive pricing and dynamic pricing strategies for the target product, appropriately adjust the price of the target product to be lower than the average price of the platform, strengthen the training of the customer service team, ensure timely response to customer inquiries, and solve post-purchase problems; optimize the return and exchange policy, reduce negative reviews caused by after-sales problems, and pay attention to consumers' evaluation and feedback on the target product. Take corrective measures to improve product quality and service for negative reviews, and then sell the target product on the e-commerce platform; When receiving the secondary sales risk level, the push information content is that the current target product is competitive in sales on the e-commerce platform and is not in a risky sales state. It is recommended that merchants sell the target product on the e-commerce platform and update and optimize the information on the product page, including high-quality pictures, product descriptions and user reviews, to increase the trust of potential buyers. Based on user search habits, the product title and keywords are described in detail to increase natural search traffic, and marketing promotion is carried out through multiple online channels to attract potential customers to pay attention to the target product.

[0037] In this embodiment, the comprehensive sales risk assessment index Zpg is calculated by associating the exposure conversion coefficient Xzh, the price fluctuation coefficient Xbd and the after-sales risk coefficient Xsh. This innovative analysis method not only improves the accuracy of risk assessment, but also provides a scientific decision-making basis for merchants. Through the hierarchical analysis method, merchants can conduct in-depth analysis of the weights of various indicators to ensure the scientificity and effectiveness of the decision. The risk rating prediction module compares the comprehensive evaluation index Zpg with the preset evaluation threshold P to judge the market competitiveness and sales risk status of the target product, predict the first and second level sales risk levels, and can push risk prompts in time to help merchants quickly adjust their sales strategies and reduce potential financial losses. For example, if the target product is assessed to be in a risky sales state, the merchant can choose to sell it after adjusting the strategy to avoid losses; when the product is assessed to be not in a risky state, the merchant can confidently sell it and increase the traffic exposure intensity of the target product, thereby improving market competitiveness. Through such risk management and decision support, merchants can not only better adapt to market changes, but also optimize resource allocation, improve overall operational efficiency, and ultimately achieve higher economic benefits.

[0038] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A risk rating prediction system based on e-commerce big data analysis, characterized by: It includes a first data collection module, a correlation analysis module, a second data collection module, a comprehensive sales risk assessment module and a risk rating prediction module; The first data collection module is used to collect the exposure frequency data of the target product in each set time period by accessing the public API interface of the e-commerce platform to obtain the exposure frequency Pbg of the target product, and collect the sales quantity data of the target product in the corresponding time period according to the sales quantity of the target product; The correlation analysis module is used to analyze the impact of the target product exposure frequency Pbg on the target product sales quantity Ssm according to the correlation between the target product exposure frequency data and the target product sales quantity data, generate the exposure correlation index Zsm, preliminarily determine whether the target product has sales competitiveness on the e-commerce platform, and issue corresponding sales risk analysis instructions; The second data collection module is used to collect the target commodity selling price information data of each period on the e-commerce platform after receiving the sales risk analysis instruction, and collect the target commodity after-sales information according to the operating conditions of the stores selling the same target commodity on the e-commerce platform, and obtain the after-sales information data of the competing stores; The comprehensive sales risk assessment module is used to obtain the comprehensive sales risk assessment index Zpg by fitting based on the target product exposure frequency data, target product sales quantity data, target product sales price information and competing store after-sales information data by using machine deep learning technology, and performing feature extraction and dimensionless processing; The risk rating prediction module is used to compare and analyze the preset evaluation threshold value P with the comprehensive sales risk evaluation index Zfx, predict the corresponding sales risk level, and make corresponding push information according to the sales risk level.

2. The risk rating prediction system based on e-commerce big data analysis according to claim 1 is characterized by: The first data acquisition module includes an access unit and a data acquisition unit; The access unit is used to register an e-commerce platform developer account, obtain an API key and document, identify available API endpoints and request formats, and use the programming language Python to construct HTTP request related parameters through a library, wherein the related parameters include a target product ID and a search keyword; The data collection unit is based on the public API interface of the connected e-commerce platform, sends a search request, receives response data, and collects the target product exposure frequency data and the target product sales quantity data in each time period after feature recognition and screening, wherein the target product exposure frequency data includes the target product exposure frequency Pbg in each time period, and the target product sales quantity data includes the target product sales quantity Ssm in each time period.

3. The risk rating prediction system based on e-commerce big data analysis according to claim 2 is characterized by: The correlation analysis module includes an exposure correlation analysis unit and a judgment unit; The exposure correlation analysis unit is used to analyze the influence of the target commodity exposure frequency Pbg on the target commodity sales quantity Ssm based on the machine learning model and in combination with the correlation between the target commodity exposure frequency data and the target commodity sales quantity data, so as to generate an exposure correlation index Zsm. The exposure correlation index Zsm is obtained by the following formula: ； In the formula, It is represented as the exposure frequency of the target product in the i-th collection period, It is represented by the sales quantity of the target product in the i-th collection period. It is expressed as the average exposure frequency of the target product during the collection period. It is represented as the mean sales quantity of the target product during the collection period, i=1, 2, 3, ..., n, where n is the collection period.

4. The risk rating prediction system based on e-commerce big data analysis according to claim 3 is characterized by: The judgment unit is used to preliminarily judge whether the target product has sales competitiveness on the e-commerce platform according to the value corresponding to the exposure correlation index Zsm, and issue a corresponding sales risk analysis instruction, the specific contents of which are as follows: If the exposure correlation index Zsm=1, it means that there is a completely positive correlation between the exposure frequency and the sales quantity of the target product. It is preliminarily judged that the target product has sales competitiveness on the e-commerce platform, and a sales risk analysis instruction is issued; If the exposure correlation index Zsm=-1, it means that there is a completely negative correlation between the exposure frequency and the sales quantity of the target product. It is preliminarily judged that the target product does not have sales competitiveness on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit; If the exposure correlation index Zsm=0, it means that there is no correlation between the exposure frequency and sales quantity of the target product. It is preliminarily judged that the target product does not have sales competitiveness on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit.

5. The risk rating prediction system based on e-commerce big data analysis according to claim 4 is characterized by: The second data collection module is used to collect target commodity selling price information data for each time period on the e-commerce platform after receiving the sales risk analysis instruction, wherein the target commodity selling price information data includes the target commodity selling price Gsm for the corresponding time period, and collect after-sales information of the target commodity based on the operating conditions of the stores selling the same target commodity on the e-commerce platform, and obtain after-sales information data of competing stores, wherein the after-sales information data of competing stores includes the target commodity return rate Vth and the target commodity negative review rate Vcp.

6. The risk rating prediction system based on e-commerce big data analysis according to claim 5 is characterized by: The comprehensive sales risk assessment module includes an exposure conversion analysis unit, a price fluctuation analysis unit, a post-sales analysis unit and a sales risk assessment unit; The exposure conversion analysis unit is based on the exposure frequency data of the target product and the sales quantity data of the target product. When the exposure correlation index Zsm=1, there is a completely positive correlation between the exposure frequency and the sales quantity of the target product. After dimensionless processing, the conversion effect between the exposure frequency and the sales quantity of the target product in each time period is analyzed to obtain the exposure conversion coefficient Xzh. The exposure conversion coefficient Xzh is obtained by the following formula; ； In the formula, It is represented as the exposure frequency of the target product in the i-th collection period, It is represented by the sales quantity of the target product in the i-th collection period, i=1, 2, 3, ..., n, and n represents the collection cycle.

7. The risk rating prediction system based on e-commerce big data analysis according to claim 6 is characterized by: The price fluctuation analysis unit is used to summarize the target commodity selling price information data to construct a target commodity selling price data set, and extract the highest target commodity selling price Gsm Max , through the statistical mean algorithm, obtain the average target product selling price in the set time period , and after linear normalization, the price fluctuation coefficient Xbd is obtained by fitting. The price fluctuation coefficient Xbd is obtained by the following formula: ； In the formula, i=1, 2, 3, ..., n, n represents the acquisition period, It is represented as the highest selling price of the target product in the i-th collection time period.

8. The risk rating prediction system based on e-commerce big data analysis according to claim 6 is characterized by: The after-sales analysis unit is used to analyze the after-sales information data of the competing stores, and obtain the after-sales risk coefficient Xsh after dimensionless processing. The after-sales risk coefficient Xsh is obtained by the following formula; ； In the formula, in the formula, It is represented as the target product return rate of the j-th store, It is represented as the negative review rate of the target product in the jth store, j=1, 2, 3, ..., m, where m is the number of stores. and are all expressed as weight values.

9. The risk rating prediction system based on e-commerce big data analysis according to claim 8 is characterized by: The sales risk assessment unit is used to utilize machine deep learning technology, and dimensionlessly process the exposure conversion coefficient Xzh, the price fluctuation coefficient Xbd, and the after-sales risk coefficient Xsh, and obtain a comprehensive sales risk assessment index Zpg by fitting. The comprehensive sales risk assessment index Zpg is obtained by the following formula: ； In the formula, , and They are respectively represented as the weight values ​​of exposure conversion coefficient Xzh, price fluctuation coefficient Xbd and after-sales risk coefficient Xsh, and A is represented as the correction constant.

10. The risk rating prediction system based on e-commerce big data analysis according to claim 9 is characterized by: The risk rating prediction module includes a rating prediction unit and a push unit; The rating prediction unit is used to compare and analyze the preset evaluation threshold value P with the comprehensive sales risk evaluation index Zpg, predict the corresponding sales risk level, and make corresponding push prompts according to the sales risk level. The specific contents are as follows; If the comprehensive sales risk assessment index Zpg ≥ the assessment threshold P, it means that although the target product has sales competitiveness on the e-commerce platform, the target product is in a risky sales state on the e-commerce platform, and the first-level sales risk level is obtained and sent to the push unit; If the comprehensive sales risk assessment index Zpg is less than the assessment threshold P, it indicates that the target product has sales competitiveness on the e-commerce platform, and the target product is not in a risky sales state on the e-commerce platform, and the secondary sales risk level is obtained and sent to the push unit; The push unit is used to receive the primary selling risk level and the secondary selling risk level, and send the corresponding push information to the display screen, the specific content is as follows; When receiving the first-level sales risk level, the push information content is that the current target product is in a risky sales state. It is recommended that the merchant first implement competitive pricing and dynamic pricing strategies for the target product, appropriately adjust the price of the target product to be lower than the average price of the platform, ensure timely response to customer inquiries, and solve post-purchase problems; at the same time, pay attention to consumers' evaluation and feedback on the target product, take corrective measures to improve product quality and service for negative reviews, and then sell the target product on the e-commerce platform; When receiving the second-level sales risk level, the pushed information content is that the current target product is competitive in sales on the e-commerce platform and is not in a risky sales state. It is recommended that merchants sell the target product on the e-commerce platform. At the same time, the information on the product page is updated and optimized, including high-quality pictures, product descriptions and user reviews. According to user search habits, the product title and keywords are described in detail to increase natural search traffic, and marketing promotion is carried out through multiple online channels to attract potential customers to pay attention to the target product.

Images