Search

CN-121981752-A - Hot-sale bid mining method based on bid price of AI intelligent marketing

CN121981752ACN 121981752 ACN121981752 ACN 121981752ACN-121981752-A

Abstract

The invention belongs to the technical field of bid mining, and discloses a hot-sale bid mining method based on bid price of AI intelligent marketing, which comprises the following steps of acquiring bid data from a plurality of data sources by using a network data acquisition technology and data interface call; the method comprises the steps of obtaining multi-source bid data, cleaning and converting the obtained multi-source bid data, loading the cleaned and converted data into a database or a data warehouse, carrying out clustering analysis on the bid data by using a clustering algorithm, comprehensively considering the heat, price competitiveness, sales and praise factors of the bid, calculating comprehensive scores for each bid, and generating a bid analysis report according to data analysis and calculation results. The hot-sale bidding product mining method has the advantages of showing remarkable advantages in accuracy, efficiency and comprehensiveness, helping enterprises to grasp market trends more accurately, optimizing products and marketing strategies, and effectively improving market competitiveness.

Inventors

  • YU LINYI
  • LI XIAO

Assignees

  • 北京森博明德人工智能科技股份有限公司

Dates

Publication Date
20260505
Application Date
20260123

Claims (10)

  1. 1. The hot sale bidding mining method based on the price of the item for AI intelligent marketing is characterized by comprising the following steps: step S1, acquiring bid data from a plurality of data sources by using a network data acquisition technology and a data interface call; S2, cleaning and converting the obtained multi-source bidding data; step S3, the cleaned and converted data are loaded into a database or a data warehouse; s4, performing cluster analysis on the bid data by using a cluster algorithm; S5, comprehensively considering the heat, price competitiveness, sales and public praise factors of the products, and calculating comprehensive scores for each bid product; And S6, generating a bid analysis report according to the data analysis and calculation results.
  2. 2. The method for mining hot-sale bidding products based on the price of the products for AI intelligent marketing according to claim 1, wherein in step S1, bidding product data is acquired from a plurality of data sources by using a network data acquisition technology and a data interface call, specifically comprising: (1) In the aspect of an e-commerce platform, the basic information, price information, sales data and user evaluation data of a product are acquired through simulating user behaviors, setting request frequency and random request heads and through respective open API interfaces or data acquisition programs; the basic information of the product comprises a product name, a brand, a model and a specification, price information comprises original price, promotion price and historical price fluctuation, sales data comprises sales volume, sales amount and sales ranking, and user evaluation data comprises evaluation content, score and evaluation time; (2) For an industry information website, acquiring an industry report, an expert viewpoint and a market analysis article by using a network data acquisition technology, and extracting information about the latest dynamics, market share change and technical innovation of the bid; (3) On a social media platform, data of discussion topics, emotion tendencies and public praise of the bid by the user are collected through related API interfaces, content, forwarding quantity and comment quantity data published by the user are obtained through searching specific bid keywords, and the attention degree and attitude of the user to the bid are analyzed.
  3. 3. The method for mining the hot-sale bidding products based on the price of the products by the intelligent marketing of the AI according to claim 1 is characterized in that in the step S2, the obtained multi-source bidding product data is cleaned and converted, and the specific process is as follows: step S21, cleaning the collected original data to remove repeated data, invalid data and error data; firstly, through writing a data cleaning script, identifying and deleting repeated product records according to the characteristics and business rules of data, and processing invalid data and correcting error data; Step S22, data are standardized and converted, and data in different formats are unified into a format convenient for analysis; and S23, preprocessing operations of word segmentation, word stopping and stem extraction are performed on the text data.
  4. 4. The method for mining hot offset bidding products based on the price of the products for AI intelligent marketing according to claim 3, wherein in step S3, the cleaned and converted data is loaded into a database or a data warehouse, specifically comprising: The method comprises the steps of selecting a database management system for storing structured data, combining a Hadoop distributed file system with a Hive data warehouse for storing unstructured data, and realizing management and query of the unstructured data by utilizing a table structure and a query language of Hive; In the data loading process, data are extracted from a data source through an ETL tool, and are loaded to corresponding storage positions according to the table structure and the storage rule of a database or a data warehouse after being cleaned and converted, and meanwhile, data indexes are built, the data storage structure is optimized, and the data query and analysis efficiency is improved.
  5. 5. The method for mining hot-sale bidding products based on the price of the products by the intelligent AI marketing method according to claim 1, wherein in the step S4, clustering analysis is performed on bidding product data by using a clustering algorithm, and the specific process is as follows: Step S41, based on a K-Means algorithm, firstly determining the number K of clusters, calculating the distance between each data point and each cluster center according to the product category, price, sales volume and characteristic data of multiple dimensions evaluated by a user, and distributing the data points to the clusters closest to each cluster; Step S42, mining potential association relations among the bid products by adopting an association rule mining algorithm, setting a minimum support degree and a minimum confidence degree threshold value based on an Apriori algorithm, generating a frequent item set, namely an item set meeting the minimum support degree, by scanning a data set, and then generating an association rule, namely a rule meeting the minimum confidence degree according to the frequent item set.
  6. 6. The method for mining the hot bid price based on the price of the product by the intelligent AI marketing method according to claim 1, wherein in the step S5, the comprehensive score is calculated for each bid by comprehensively considering the factors of the heat, price competitiveness, sales volume and public praise of the product, and the concrete process is as follows: Step S51, based on a class heat calculation mode, collecting data of search quantity, sales quantity, attention number and topic discussion quantity of a certain class, and calculating a heat value of the class by combining a predetermined weight coefficient, wherein the method specifically comprises the following steps: firstly, collecting original data of search quantity, sales quantity, discussion quantity and negative comment quantity of a target class in the last 30 days; then, carrying out normalization processing on each piece of original data, mapping to a 0-100 interval, and eliminating dimension differences; finally, substituting the weighted coefficient to calculate a weighted sum to obtain a class heat value; Step S52, calculating the price competitiveness of the product relative to the bid according to the price competitiveness mode; firstly, determining the average price of the product, namely, the trading average price of all the competing products of the product in the last 30 days; secondly, calculating a price relative coefficient of the target bid product according to the ratio of the bid product price to the product average price; then, the quality/public praise score is extracted through AI text analysis comments, and the cost performance index of the scoring bid product is calculated by weighting the product configuration, quality and public praise index in combination with the quantization of the product configuration parameters; Finally, substituting Sigmoid function to calculate price competitiveness The nonlinear normalization of competitiveness is realized; step S53, calculating a comprehensive score for each bid according to a comprehensive score calculation mode by comprehensively considering heat, price competitiveness, sales and public praise factors of the bid; firstly, based on the calculation result, obtaining the heat degree of the product Competitive price ; Secondly, calculating sales volume increase rate and market share of the bid, and normalizing to a 0-100 interval respectively; then substituting the comprehensive weights to calculate a weighted sum to obtain a bid comprehensive score; finally, setting a threshold value and screening out Is the target hot-sell bidding.
  7. 7. The method for mining the hot bid amount of the intelligent AI marketing based on the price of the bid amount in step S51 is characterized in that the calculation formula of the heat value of the bid amount is as follows: ; Wherein, the The value range of the comprehensive heat value is 0-100, and the higher the value is, the stronger the heat value is; 、 、 And Are all weight coefficients, satisfy And in particular, the method comprises the steps of, In order for the search weights to be of a type, Is sales weight, For the purposes of discussion of the weight, The weight coefficient can be dynamically adjusted according to the industry characteristic; The heat standard value is searched for the category, and is obtained by normalizing the searching amount of the whole network in the last 30 days; the standard value of the heat degree of the sales quantity is calculated by the average sales quantity of the sales quantity in the last 30 days; For the standard value of the heat degree of the class discussion, counting the related topic quantity/comment quantity of social media and electronic commerce comment areas; the standardized value of the negative rate of the category is calculated by the negative comment number/total comment number, and specifically, 、 、 And The calculation formulas of (a) are respectively as follows: ; ; ; 。
  8. 8. The method for mining hot-sale bidding products based on the price of the product class for AI intelligent marketing as claimed in claim 6, wherein in step S52, the price relative coefficient is The calculation formula of (2) is as follows: ; Wherein, the Is a bid price relative coefficient, when the bid price is 2 times higher than the average price of the product class, Taking 10, showing the high price disadvantage, when the price of the bid product is lower than 50 percent of the average price of the product class, 2.5, Avoiding the excessive advantage of low price; Cost performance index The calculation formula of (2) is as follows: ; Wherein, the Scoring the price competition ratio; For the index weight of the i-th item, For the score of the index of the i-th item, For the number of metrics, by default m=3, =0.4、 =0.3、 =0.3; Price competitiveness The calculation method of (2) is as follows: ; Wherein, the The competitive value of the bid product in the corresponding price segment is 0-100, and the higher the value is, the stronger the competitive force is; To adjust the coefficients, default The steepness of the competitive curve is controlled and adjusted according to industry pricing sensitivity.
  9. 9. The method for mining hot bid amount based on the bid amount of AI intelligent marketing according to claim 6, wherein in step S53, the bid amount composite score F is as follows: ; Wherein F is a bid comprehensive score, and the score F is more than or equal to 80, and is judged to be a hot bid; 、 、 And Are all comprehensive weights, satisfy And in particular, the method comprises the steps of, Is of the grade heat degree, Is competitive in price, Is the sales increase rate, Is the market share; Normalizing the value for the sales growth rate; Standardizing values for market shares; And The calculation modes of (a) are respectively as follows: ; Wherein, the When the percentage is more than or equal to 100%, taking 100, and taking 0 for negative growth; 。
  10. 10. The method for mining hot-sale bidding products based on the price of the products for AI intelligent marketing according to claim 1, wherein in step S6, detailed bidding product analysis report is generated according to the data analysis and calculation result; The report content comprises basic information of the bid, market performance, price competitiveness analysis, association relation analysis and comprehensive evaluation; The basic information of the bid product comprises brands, product models, products and price ranges, the market performance comprises sales, market share and product heat ranking, the price competitiveness analysis comprises price difference, cost performance and price competitiveness ranking, the association relation analysis comprises association conditions with other bid products or products, and the comprehensive evaluation comprises comprehensive grading, advantage and disadvantage analysis and market positioning summarization; the data is presented visually in the form of tables and charts while suggestions and policies based on analysis of the data are provided in the report.

Description

Hot-sale bid mining method based on bid price of AI intelligent marketing Technical Field The invention relates to the technical field of bid mining, in particular to a hot-sale bid mining method based on a bid price of AI intelligent marketing. Background In today's highly competitive market environment, enterprises face unprecedented challenges and opportunities. With the continuous subdivision of the market and the increasing diversity of consumer demands, enterprises need to accurately understand market dynamics, and particularly, deep grasp of the bid product, has become a key factor in standing and gaining advantages in the market. The bid analysis is an important means for enterprises to formulate marketing strategies, optimize product services and improve competitiveness, and has self-evident value. Through the comprehensive analysis of the bid, the enterprises can clearly recognize the self-positioning in the market and find out the advantages and the disadvantages of the self-product or service, thereby improving and innovating pertinently. Meanwhile, the information such as marketing strategies, price systems and target customer groups of the bid products is deeply known, the enterprises are helped to formulate more targeted and differentiated marketing strategies, homogenization competition is effectively avoided, and market share and profitability are improved. The traditional hot-sale bidding mining method mainly relies on manual collection and analysis of data, and has a plurality of limitations. In the aspect of data collection, the manual operation efficiency is low, and massive market data is difficult to comprehensively cover. For example, on an electronic commerce platform, the variety of commodities is various, the updating speed is high, and the bidding information is screened and collected one by means of manpower, so that a great deal of manpower and time are consumed, and important potential bidding products are easily omitted. Moreover, the accuracy and integrity of manually collected data is difficult to guarantee and may be limited by subjective factors and information acquisition channels. In the data processing and analysis stage, manual analysis is difficult to cope with complex data relationships and large-scale data volumes, and valuable information cannot be extracted quickly and accurately. The traditional method can only analyze a limited number of dimensions, and can not mine the characteristics and advantages of the bid from multiple dimensions and deep layers, so that the analysis result is not comprehensive and deep enough, and sufficient support can not be provided for enterprise decision-making. In the face of rapid changes of markets, the information updating of the traditional method is lagged, so that the latest conditions of market dynamics and bidding products are difficult to reflect in time, and enterprises are in a passive status in competition. Disclosure of Invention The invention aims to provide a hot sale bidding mining method based on the price of a bid by AI intelligent marketing, which realizes comprehensive monitoring and accurate positioning of the bid by collecting and analyzing data of multiple platforms and multiple channels in real time, and utilizes a machine learning algorithm and a big data analysis technology, AI can conduct deep analysis on the bid from multiple dimensions, including product characteristics, price strategies, sales channels, user evaluation and the like, so as to provide more comprehensive and deep bidding insight for enterprises and help the enterprises to preempt in vigorous market competition. In order to achieve the above purpose, the invention provides a hot-sale bid mining method based on the bid price of AI intelligent marketing, which comprises the following steps: step S1, acquiring bid data from a plurality of data sources by using a network data acquisition technology and a data interface call; S2, cleaning and converting the obtained multi-source bidding data; step S3, the cleaned and converted data are loaded into a database or a data warehouse; s4, performing cluster analysis on the bid data by using a cluster algorithm; S5, comprehensively considering the heat, price competitiveness, sales and public praise factors of the products, and calculating comprehensive scores for each bid product; And S6, generating a bid analysis report according to the data analysis and calculation results. Preferably, in step S1, the bid data is acquired from a plurality of data sources by using a network data acquisition technology and a data interface call, and specifically includes: (1) In the aspect of an e-commerce platform, the basic information, price information, sales data and user evaluation data of a product are acquired through simulating user behaviors, setting request frequency and random request heads and through respective open API interfaces or data acquisition programs; the basic information of the product compri