A method for automatically pushing orders based on system connection based on big data

Abstract

The present invention relates to the field of automatic order promotion technology, specifically a system docking automatic order promotion method based on big data, comprising the following steps: accessing news source websites, collecting news text content, extracting news event information, matching the news event information with known types, and generating event category classification results. In the present invention, the application of consumer behavior analysis results not only improves user experience, but also predicts popular product bundles and optimizes inventory management. The dynamic adjustment of price points combined with consumer price sensitivity data effectively improves sales conversion rates and reduces marketing costs.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic order pushing, and in particular to a system-connected automatic order pushing method based on big data. Background Art

[0002] Automated order processing is a key component of business process automation, with a particular focus on the automated processing of online purchase, sale or rental transactions. The technology primarily involves the use of computer systems and software tools to automate order processing, thereby increasing efficiency and reducing the need for human intervention. Automated order processing includes electronic shopping and electronic order processing, enabling businesses to process large volumes of transactions quickly and accurately.

[0003] However, existing technologies fail to effectively integrate external real-time information, such as news events, resulting in delayed responses when the market changes rapidly. This lag may result in failure to adjust product recommendation strategies in a timely manner, thereby missing out on key sales opportunities. For example, when faced with emergencies, existing technologies are unable to quickly reflect the impact of the information on consumer behavior, which in turn affects product sales and customer satisfaction, and reduces potential purchase conversion rates. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a method for automatically pushing orders through system docking based on big data.

[0005] In order to achieve the above object, the present invention adopts the following technical solution, a method for automatically pushing orders based on system docking based on big data, comprising the following steps:

[0006] Visit news source websites, collect news text content, extract news event information, match the news event information with known types, and generate event category classification results;

[0007] Based on the event classification results, extract the category identifier of each event, calculate the correlation between each news event category and each product, generate a correlation analysis result, adjust the display priority of the product recommendation according to the correlation analysis result, dynamically update the product recommendation list according to the display priority, and complete the recommendation list update;

[0008] Perform data aggregation based on consumer purchase records and the association analysis results to generate consumer behavior analysis results, use the consumer behavior analysis results to calculate potential bundling sales opportunities between commodities, predict popular product bundles, and generate bundling sales optimization solutions;

[0009] Based on the bundled sales optimization plan, perform price elasticity analysis, calculate the price adjustment point of each bundled product combination, perform dynamic adjustment of price points, adjust the sales price of bundled products based on the price sensitivity data of each consumer group, generate a price adjustment strategy, use the price adjustment strategy to make personalized discount recommendations, analyze the acceptance rate and effect of discount recommendations, and obtain personalized discount recommendation results.

[0010] Preferably, the steps of obtaining the event classification result are:

[0011] Visit the news source website, collect news text content, and obtain the original text data set;

[0012] Based on the original text dataset, entity recognition in natural language processing is performed to identify the key information of the news event. The formula is:

[0013] ;

[0014] in, For key information of news events, The original text News terms, is the weight of the term, is the total number of terms;

[0015] Based on the key information of the news event, it is matched with a preset event type database, classified according to the matching degree, and the event category classification result is obtained.

[0016] Preferably, the steps of obtaining the association analysis result are:

[0017] According to the event classification result, extract the category identifier of each event to obtain an event category identifier set;

[0018] Based on the event category identification set, a correlation analysis is performed to calculate the correlation between each news event category and each commodity. The calculation formula is:

[0019] ;

[0020] in, For event category With goods The correlation degree, For Category Next The frequency of occurrence of an event, For events With goods The prior strength of association, For Category The total number of events in ;

[0021] Based on the degree of association, an association analysis result is generated.

[0022] Preferably, the steps of updating the recommendation list are:

[0023] Using the correlation analysis results, sort each product according to its correlation with the current news event to obtain a product recommendation priority list;

[0024] Based on the commodity recommendation priority list, real-time monitoring of user activities and market changes is performed to adjust commodity recommendation priorities in real time;

[0025] The adjusted product recommendation priority is applied to online display, and the product recommendation list displayed on the front end is dynamically updated according to the adjusted product recommendation priority to complete the update of the recommendation list.

[0026] Preferably, the steps for obtaining the consumer behavior analysis results are:

[0027] Collecting the consumer's purchase record data, wherein the purchase record data includes purchase frequency, purchase time, type of goods purchased, and quantity of goods purchased, to obtain a consumer purchase record set;

[0028] In combination with the association analysis results, the consumer purchase record set is analyzed to match the consumer purchase record with the association between the current product and event, and to identify the consumer's purchase preference and behavior pattern;

[0029] Perform data aggregation to generate consumer behavior analysis results based on purchasing preferences and behavior patterns.

[0030] Preferably, the steps for obtaining the bundled sales optimization solution are:

[0031] Extract key data from the consumer behavior analysis results, wherein the key data includes the consumer's interest level and purchase tendency for each category of goods, to form a consumer interest and purchase tendency data set;

[0032] Based on the consumer interest and purchase tendency dataset, the potential bundling opportunities between products are calculated using the following formula:

[0033] ;

[0034] in, Display products and products The potential value of bundling sales, For goods According to consumers The purchase intention score is obtained. For goods The score obtained based on the purchase tendency of the same consumer, For consumers Purchase activity;

[0035] According to the bundling sales potential value, a product combination is determined and a bundling sales optimization plan is generated.

[0036] Preferably, the steps of obtaining the price adjustment strategy are:

[0037] Extracting the potential sales value of each commodity bundle combination from the bundle sales optimization scheme to form a bundle commodity value data set;

[0038] Based on the price sensitivity data of each consumer group and the bundled product value data set, a price elasticity analysis is performed to calculate the price adjustment point of each bundled product combination. The calculation formula is:

[0039] ;

[0040] in, represents the adjusted price, As the basic starting price, For consumer groups The change in demand, is the price elasticity of the consumer group, is the total number of consumer groups;

[0041] According to the adjusted price, the sales price of each bundled product is adjusted to generate a price adjustment strategy.

[0042] Preferably, the steps for obtaining the personalized preferential recommendation results are:

[0043] Based on the price adjustment strategy, the discount content is customized according to each consumer's purchase record, behavior pattern and corresponding price sensitivity to form personalized discount recommendations;

[0044] Implement personalized offer recommendations to consumer groups via email, social media, and in-app notifications;

[0045] Collect and analyze consumers' acceptance rate of the personalized discount recommendations, determine the effectiveness of the discount implementation, and generate personalized discount recommendation results.

[0046] Compared with the prior art, the advantages and positive effects of the present invention are:

[0047] The present invention obtains text data from news sources and implements entity recognition to extract event information, so that product recommendations can be adjusted based on the latest market dynamics, thereby optimizing timeliness and relevance. The classification and analysis of event information improves the accuracy of product recommendations and enhances the synchronization of recommended content with market trends. The correlation between news event categories and products is further analyzed, and the recommendation list is dynamically updated to make the product display closer to consumers' current needs and interests. The application of consumer behavior analysis results not only improves user experience, but also predicts popular product bundles and optimizes inventory management. The dynamic adjustment of price points combined with consumers' price sensitivity data improves sales conversion rates and reduces marketing costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0050] See also Figure 1 The present invention provides a technical solution, a method for automatically pushing orders based on system docking of big data, comprising the following steps:

[0051] Visit news source websites, collect news text content, extract news event information, match the news event information with known types, and generate event category classification results;

[0052] Based on the event classification results, extract the category identifier of each event, calculate the correlation between each news event category and each product, generate the correlation analysis results, adjust the display priority of product recommendations based on the correlation analysis results, dynamically update the product recommendation list based on the display priority, and complete the recommendation list update;

[0053] Perform data aggregation based on consumer purchase records and association analysis results to generate consumer behavior analysis results. Use the consumer behavior analysis results to calculate potential bundling sales opportunities between products, predict popular product bundles, and generate bundling sales optimization plans;

[0054] Based on the bundled sales optimization plan, perform price elasticity analysis, calculate the price adjustment point of each bundled product combination, perform dynamic adjustment of price points, adjust the sales price of bundled products based on the price sensitivity data of each consumer group, generate a price adjustment strategy, use the price adjustment strategy to make personalized discount recommendations, analyze the acceptance rate and effect of discount recommendations, and obtain personalized discount recommendation results.

[0055] Specifically, by obtaining text data from news sources and implementing entity recognition to extract event information, product recommendations can be adjusted based on the latest market dynamics, thereby optimizing the timeliness and relevance of the recommendation system. The classification and analysis of event information improves the accuracy of product recommendations and enhances the synchronization of recommended content with market trends. Further analysis is made of the correlation between news event categories and products, and the recommendation list is dynamically updated to make product displays more closely aligned with consumers' current needs and interests. The application of consumer behavior analysis results not only improves user experience, but also accurately predicts popular product bundles and optimizes inventory management. The dynamic adjustment of price points combined with consumer price sensitivity data effectively improves sales conversion rates and reduces marketing costs.

[0056] The steps to obtain the event classification results are as follows:

[0057] Visit the news source website, collect news text content, and obtain the original text data set;

[0058] Based on the original text dataset, we perform entity recognition in natural language processing to identify the key information of news events. The formula is:

[0059] ;

[0060] in, For key information of news events, The original text News terms, is the weight of the term, is the total number of terms;

[0061] Based on the key information of the news event, it is matched with the preset event type database, classified according to the matching degree, and the event category classification result is obtained.

[0062] Specifically, the formula The benefits of this method are that it can more accurately extract the main content related to online shopping in news texts, such as product releases, sales activities or consumer trends, etc. It can effectively grasp the main idea of ​​news texts, improve the accuracy of information classification and directly connect to the product recommendation system.

[0063] Calculation process: Setting weight parameters is the term frequency-inverse document frequency (TF-IDF) value, each term The frequency of a word is inversely correlated with the frequency of documents in the news database. For example, if a news text mentions "discount clearance", the weight of the term "discount" is determined by the inverse proportion of its frequency of occurrence in all texts. By adjusting the weight of each word, the contribution of keywords to the meaning of the text can be highlighted.

[0064] If the TF-IDF value of the word "discount" is 0.08 and it appears 5 times, then , , .

[0065] The results show that A high value indicates that the word “discount” contributes a lot to the key information of the news event, and the news is likely to be related to online shopping promotions.

[0066] By accessing specific news sources, we collect text data from different reports. These data contain various characters and formats without being processed, and need to be initially cleaned and standardized to ensure the accuracy and efficiency of subsequent analysis. The collected raw data mainly includes the title, release time and content of the news. By regularly updating the database, we keep the data fresh and relevant. This step ensures that we obtain enough samples for in-depth analysis, thereby providing basic data support for entity recognition and event classification.

[0067] The key information of the news event is compared and matched with the preset event type database. The matching degree of each type of news event with the defined event type in the database is calculated, and the matching algorithm is used to determine the closest event type. For example, if the key information is concentrated in words such as "limited-time discount" and "new product launch", such news will be classified as online shopping promotion events. The result of this step is to mark the exact event category for each news item.

[0068] The steps to obtain the association analysis results are as follows:

[0069] According to the event category classification results, the category identifier of each event is extracted to obtain an event category identifier set;

[0070] Based on the event category identification set, a correlation analysis is performed to calculate the correlation between each news event category and each product. The calculation formula is:

[0071] ;

[0072] in, For event category With goods The correlation degree, For Category Next The frequency of occurrence of an event, For events With goods The prior strength of association, For Category The total number of events in ;

[0073] Based on the correlation degree, an association analysis result is generated.

[0074] Specifically, according to the generated event category classification results, continue to analyze the obtained event category identifiers to ensure that the classification of each news event is correct, and use database query and data matching technology to extract the category identifier of each event from the database, including but not limited to text labels, digital IDs, etc., to ensure that these event category identifier sets can accurately reflect the classification attributes of the events. By setting query parameters and calling database interfaces for efficient data extraction, a clear and accurate set of event category identifiers can be obtained.

[0075] formula The benefits of this are: it can quantify the correlation between news event categories and products, provide a scientific basis for the product recommendation system, and enhance the relevance and accuracy of recommendations.

[0076] Calculation process: a specific event category Below events, the frequency of occurrence of each event Calculated by dividing the number of reports in the past week by the total number of reports, with each value averaging approximately 0.1. The strength of the prior association with each event Obtained through correlation analysis of historical sales data and event reports, the average value is set to 0.05. Substitute into the formula and calculate:

[0077] ;

[0078] The results show that: Event type With goods The correlation degree is 0.005, indicating that the correlation between the two is not strong. It is necessary to adjust the priority of the product recommendation list based on this analysis result.

[0079] Based on the results of the correlation analysis, statistical analysis and data visualization technology are used to further process the data and generate a detailed correlation analysis result table. The table not only includes the correlation value between each event category and the product, but also sorts them according to the level of correlation, providing an intuitive and easy-to-understand display format, providing a basis for subsequent product recommendation decisions, and ensuring that the recommendation system can be dynamically adjusted according to actual data, ultimately forming a product recommendation list that meets user needs and market dynamics.

[0080] The steps to update the recommendation list are:

[0081] Using the correlation analysis results, sort each product according to its correlation with the current news event to obtain a product recommendation priority list;

[0082] Based on the product recommendation priority list, monitor user activities and market changes in real time, and adjust the product recommendation priority in real time;

[0083] The adjusted product recommendation priority is applied to online display, and the product recommendation list displayed on the front end is dynamically updated according to the adjusted product recommendation priority to complete the update of the recommendation list.

[0084] Specifically, starting from the association analysis result table, the association between each product and the current news event is digitally scored, and the products are sorted in descending order according to the association value through a sorting algorithm. The association score of each product is calculated by comparing its correlation index with the current hot news event. The correlation index includes but is not limited to the frequency of mentions, event influence and product sales history data. These data are integrated and the total association score of each product is calculated through a composite function. The composite function used in this process mainly includes weighted average and normalization processing to ensure that the score fluctuates within a reasonable range, thereby ensuring that the recommendation system can adjust its recommendation strategy in a timely manner, update the product list in real time, and finally form a preliminary product recommendation priority list sorted by correlation.

[0085] Based on the product recommendation priority list, the real-time monitoring system of the back-end server is used to continuously track users' browsing history and click behavior as well as sales trends and inventory changes in the market, and implement dynamic display logic: the data processing involved in this process includes real-time data stream capture, fast data access and high-frequency updates, and the use of machine learning algorithms to dynamically adjust the product recommendation priority to ensure that the recommendation list can reflect the user's latest needs and interests in real time, and predict their possible purchasing preferences by analyzing their shopping patterns and feedback. At the same time, it responds sensitively to market trends and adjusts the order of product display, so that the recommendation system can automatically optimize the recommendation effect without human intervention.

[0086] The updated product recommendation priority is synchronized to the online display, and the front-end display of the products is updated according to the new priority. During this update process, the back-end system will dynamically adjust the product records in the database according to the output instructions of the recommendation algorithm, and the front-end display will adjust the product arrangement on the page according to the latest database status to ensure that the latest recommended products are available to users every time they visit. The implementation process uses cache update and page rendering technology to optimize data transmission and loading speed, improve user browsing experience, and ultimately generate an updated recommendation list, which directly affects users' purchasing decisions and enhances sales potential and market response speed.

[0087] The steps to obtain the results of consumer behavior analysis are:

[0088] Collecting the purchase record data of consumers, including purchase frequency, purchase time, types of goods purchased and quantity of goods purchased, to obtain a consumer purchase record set;

[0089] Combined with the results of association analysis, analyze the consumer purchase record set, match the consumer purchase record with the current product and event correlation, and identify the consumer's purchase preferences and behavior patterns;

[0090] Perform data aggregation to generate consumer behavior analysis results based on purchasing preferences and behavior patterns.

[0091] Specifically, the purchase frequency, purchase time, commodity type and quantity data of each consumer are selected in detail from the consumer purchase record set. These data are directly imported from the sales system of each merchant into the central database for unified processing. Each data point is verified by a data verification algorithm to ensure its accuracy and exclude any outliers. The consumer's purchasing habits and preference patterns are calculated based on these data. In this way, the analysis team can understand the activity level and preference changes of each consumer.

[0092] Compare consumer purchase records with the correlation between commodities and events in the current market, and use algorithms to analyze the preference intensity of different consumers for various commodities. These preference data will be used to build a multi-dimensional consumer behavior model, which not only considers the consumer's historical purchase behavior, but also takes the current market trend as an influencing factor. During this process, each piece of consumer data undergoes strict quality control to ensure the reliability of the analysis results. Through these steps, the analysis team can identify consumers' possible purchase preferences and behavior patterns in detail, and obtain more accurate consumer behavior analysis predictions.

[0093] After obtaining detailed consumer behavior patterns, data aggregation operations are performed. This operation involves integrating behavioral pattern data from different data sources and using machine learning to analyze these data to obtain information about consumers' preferences for different product categories. This analysis result will directly affect the adjustment of the product recommendation system, enabling it to display the most appropriate product recommendations for each consumer. This step ensures the personalization and accuracy of the recommendation system and generates consumer behavior analysis results, which describe in detail which product categories consumers may be interested in and predict their future purchasing behavior.

[0094] The steps to obtain the bundled sales optimization plan are:

[0095] Extract key data from the results of consumer behavior analysis, including consumers' interest level and purchasing tendency for each category of goods, to form a consumer interest and purchasing tendency data set;

[0096] Based on the consumer interest and purchase tendency dataset, the potential bundling opportunities between products are calculated using the following formula:

[0097] ;

[0098] in, Display products and products The potential value of bundling sales, For goods According to consumers The purchase intention score is obtained. For goods The score obtained based on the purchase tendency of the same consumer, For consumers Purchase activity;

[0099] Determine the product combination based on the bundling sales potential value and generate a bundling sales optimization plan.

[0100] Specifically, by extracting key data from the consumer behavior analysis results obtained in the previous step, which include consumers' interest levels in various commodities and their purchasing tendencies, a consumer interest and purchasing tendency data set is formed. This data set reflects consumers' diverse needs and behavior patterns. By aggregating consumers' historical purchase data, including frequency, type, purchase time, etc., each consumer's interest level in various commodities and their purchasing tendencies are calculated. These data can be used to more accurately predict consumers' future purchasing behavior and provide support for subsequent marketing strategies and product recommendations. The data set aggregates these segmented information to form a specific consumer portrait. This process involves complex data processing and analysis, and the accuracy and real-time nature of the data must be ensured so that the latest purchasing tendencies of consumers can be reflected in real time.

[0101] The benefit of the formula is that by calculating the correlation between each product and the consumer's purchasing tendency, it can accurately predict and recommend product bundles that consumers may be interested in, optimize the relevance and attractiveness of product recommendations, enhance the consumer's purchasing decision-making process, and thus improve sales conversion rate; parameter For goods and products The bundling potential is obtained by calculating the correlation between products; parameter For goods According to consumers The purchase tendency score is obtained by analyzing the consumer's historical purchase data and purchase behavior patterns; parameters For goods According to consumers The purchase tendency score is also obtained by analyzing the consumer's historical purchase data and purchase behavior patterns; parameter For consumers The purchasing activity is obtained by counting the purchase frequency and interaction data of consumers within a certain period of time.

[0102] Calculation process: Assuming that consumer A's purchase propensity scores for products X and Y are 2.0 and 3.0 respectively, and the score for product Z is 1.5, the bundled sales potential of products X and Y is calculated as: (2.0×3.0+2.0×1.5) / 3=3, indicating that the bundle of products X and Y has a higher market sales potential. The result shows that according to consumer A's purchase propensity score, the bundled sales potential value of products X and Y is 3, indicating that this bundled combination may be more popular with consumers and has important reference value for sales strategies.

[0103] Based on the calculated bundling sales potential value, the most promising product combination is determined, and a specific bundling sales optimization plan is generated, which details which product combinations can be recommended as bundled sales packages. The plan ranks products based on their bundling sales potential values ​​and selects product combinations with high potential values ​​for recommendation. The optimization plan includes strategic display and promotion of bundled combinations of potential hot-selling products to attract consumers' attention and purchasing intention.

[0104] The steps to obtain the price adjustment strategy are:

[0105] Extract the potential sales value of each product bundle combination from the bundle sales optimization plan to form a bundle product value data set;

[0106] Based on the price sensitivity data of each consumer group and the bundled product value data set, a price elasticity analysis is performed to calculate the price adjustment point of each bundled product combination. The calculation formula is:

[0107] ;

[0108] in, represents the adjusted price, As the basic starting price, For consumer groups The change in demand, is the price elasticity of the consumer group, is the total number of consumer groups;

[0109] According to the adjusted price, the sales price of each bundled product is adjusted to generate a price adjustment strategy.

[0110] Specifically, the potential sales value of each commodity bundle combination is extracted from the bundle sales optimization plan to form a bundled commodity value data set. By analyzing the existing sales data and market feedback, the current market acceptance of each commodity combination and its potential for improvement are determined. Data processing includes but is not limited to consumer purchasing behavior analysis, competitor sales strategy analysis, and price sensitivity testing. All these data are cleaned and pre-processed to ensure data accuracy and reliability. Statistical analysis methods such as cluster analysis are used to determine the performance differences of each commodity combination in different consumer groups, thereby forming a potential sales value data set for each commodity bundle combination. This provides a scientific basis for subsequent price elasticity analysis and optimal price adjustment.

[0111] formula The benefit of is that by combining the demand changes and price elasticity of consumer groups, the sales price of each commodity bundle can be dynamically adjusted to achieve the best market response and maximize revenue; parameter The basic starting price is determined through market analysis, usually based on past sales data and market competition. The change in demand obtained through analysis of consumer purchase data, usually calculated based on historical sales data and market research results; parameter is the price elasticity of the consumer group obtained through price elasticity testing, which is usually obtained through experimental design or historical data regression analysis; parameter The total number of consumer groups identified through market research, usually defined in terms of market segments and target groups.

[0112] Calculation process: First, obtain the change in demand for a specific product combination for each consumer group through market research and historical sales data analysis. and price elasticity , for example, considering the base starting price It is set based on market conditions and cost analysis, for example, 500 yuan. The change in demand for a certain product bundle by a consumer group is 50 units, and the price elasticity is 0.8, which indicates that the price of this product combination is more sensitive to demand; calculate the optimal price adjustment point :

[0113] ;

[0114] This calculation shows that the adjusted optimal price is 437.5 yuan. This adjustment reflects a more delicate and moderate price change, aimed at balancing market demand and maximizing revenue, while taking into account consumers' price sensitivity.

[0115] According to the optimal price, the sales price of each bundled product is adjusted to generate a price adjustment strategy. This process involves a systematic review and repositioning of the sales price of each bundled product combination, including market competition analysis, consumer feedback collection, cost analysis, and expected profit calculation. By using pricing software or pricing models, combined with the company's pricing policy and marketing strategy, the new sales price of each product combination is determined. The generation of price adjustment strategies not only reflects the dynamic changes in the market, but also takes into account consumers' acceptance and purchasing power, ensuring that each product combination can remain competitive in the market and achieve sales targets.

[0116] The steps to obtain personalized discount recommendation results are as follows:

[0117] Based on the price adjustment strategy, the discount content is customized according to each consumer's purchase history, behavior pattern and price sensitivity, forming personalized discount recommendations;

[0118] Implement personalized offer recommendations to consumer groups via email, social media, and in-app notifications;

[0119] Collect and analyze consumers’ acceptance rate of personalized discount recommendations, judge the effectiveness of discount implementation, and generate personalized discount recommendation results.

[0120] Specifically, based on the price adjustment strategy, detailed personalized discount recommendation plans are formulated for different consumer groups. First, based on each consumer's historical purchase data and behavior patterns, a consumer portrait is synthesized using data aggregation methods. The portrait includes but is not limited to purchase frequency, preferred categories, and average consumption amount. Then, the consumer's sensitivity to price changes is analyzed, and consumers are divided into three categories of high, medium, and low sensitivity through a classification algorithm. Each type of consumer corresponds to a different discount strength according to its sensitivity. For example, for highly sensitive consumers, a larger discount is provided to attract their consumption, and for low-sensitivity consumers, customized product bundling or small discounts are provided. The data analysis and processing tools used in this process must ensure the accuracy and real-time nature of the data to ensure the personalization and effectiveness of the discount recommendation. The personalized discount recommendation data set formed according to these strategies can be directly applied to subsequent marketing activities.

[0121] Based on the generated personalized discount recommendation data set, personalized discount recommendations are implemented. First, discount information is released through multiple channels such as email, social media, and in-app notifications to ensure that all target consumer groups are covered. Second, consumers’ click and feedback data on the pushed discount information is collected in real time, and consumers’ response speed and preferences for various discounts are analyzed. The discount depth and push frequency are adjusted to enhance the pertinence and attractiveness of the recommendations.

[0122] After the personalized preferential recommendation is executed, relevant response data is collected for analysis. These data include but are not limited to the usage of coupons, purchase conversion rate and customer satisfaction survey results. First, set up data collection points to track the usage of each coupon and related sales data. Then, use statistical analysis tools to calculate the acceptance rate and sales improvement effect of preferential recommendations, aggregate and analyze the data, identify which preferential measures are most effective and which need improvement, and finally generate a personalized preferential recommendation result report that includes detailed data analysis and recommendation effect evaluation. The report will clearly point out the cost-effectiveness and improvement direction of various preferential measures to ensure the continuous optimization of marketing activities and maximize the effect.

[0123] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for automatically pushing orders through system docking based on big data, characterized in that: The following steps are involved: Visit news source websites, collect news text content, extract news event information, match the news event information with known types, and generate event category classification results; Based on the event classification results, extract the category identifier of each event, calculate the correlation between each news event category and each product, generate a correlation analysis result, adjust the display priority of the product recommendation according to the correlation analysis result, dynamically update the product recommendation list according to the display priority, and complete the recommendation list update; Perform data aggregation based on consumer purchase records and the association analysis results to generate consumer behavior analysis results, and use the consumer behavior analysis results to calculate potential bundling sales opportunities between commodities, predict popular product bundles, and generate bundling sales optimization solutions; Based on the bundled sales optimization plan, perform price elasticity analysis, calculate the price adjustment point of each bundled product combination, perform dynamic adjustment of price points, adjust the sales price of bundled products based on the price sensitivity data of each consumer group, generate a price adjustment strategy, use the price adjustment strategy to make personalized discount recommendations, analyze the acceptance rate and effect of discount recommendations, and obtain personalized discount recommendation results.

2. The method for automatically pushing orders based on system connection based on big data according to claim 1 is characterized in that: The steps for obtaining the event classification result are as follows: Visit the news source website, collect news text content, and obtain the original text data set; Based on the original text dataset, entity recognition in natural language processing is performed to identify the key information of the news event. The formula is: ； in, For key information of news events, The original text News terms, is the weight of the term, is the total number of terms; Based on the key information of the news event, it is matched with a preset event type database, classified according to the matching degree, and the event category classification result is obtained.

3. The method for automatically pushing orders based on system connection based on big data according to claim 1 is characterized in that: The steps for obtaining the association analysis results are: According to the event classification result, extract the category identifier of each event to obtain an event category identifier set; Based on the event category identification set, a correlation analysis is performed to calculate the correlation between each news event category and each commodity. The calculation formula is: ； in, For event category With goods The correlation degree, For event category Next The frequency of occurrence of an event, For events With goods The prior strength of association, For event category The total number of events in ; Based on the degree of association, an association analysis result is generated.

4. The method for automatically pushing orders based on system connection based on big data according to claim 1 is characterized in that: The steps of updating the recommendation list are: Using the correlation analysis results, sort each product according to its correlation with the current news event to obtain a product recommendation priority list; Based on the commodity recommendation priority list, real-time monitoring of user activities and market changes is performed to adjust commodity recommendation priorities in real time; The adjusted product recommendation priority is applied to online display, and the product recommendation list displayed on the front end is dynamically updated according to the adjusted product recommendation priority to complete the update of the recommendation list.

5. The method for automatically pushing orders based on system connection based on big data according to claim 1 is characterized in that: The steps for obtaining the consumer behavior analysis results are: Collecting the consumer's purchase record data, wherein the purchase record data includes purchase frequency, purchase time, type of goods purchased, and quantity of goods purchased, to obtain a consumer purchase record set; In combination with the association analysis results, the consumer purchase record set is analyzed to match the consumer purchase record with the association between the current product and event, and to identify the consumer's purchase preference and behavior pattern; Perform data aggregation to generate consumer behavior analysis results based on purchasing preferences and behavior patterns.

6. The method for automatically pushing orders based on system connection based on big data according to claim 1, characterized in that: The steps for obtaining the bundling sales optimization solution are as follows: Extract key data from the consumer behavior analysis results, wherein the key data includes the consumer's interest level and purchase tendency for each category of goods, to form a consumer interest and purchase tendency data set; Based on the consumer interest and purchase tendency dataset, the potential bundling opportunities between products are calculated using the following formula: ； in, Display products and products The potential value of bundling sales, For goods According to consumers The purchase intention score is obtained. For goods The score obtained based on the purchase tendency of the same consumer, For consumers Purchase activity; According to the bundling sales potential value, a product combination is determined and a bundling sales optimization plan is generated.

7. The method for automatically pushing orders based on system connection based on big data according to claim 1, characterized in that: The steps for obtaining the price adjustment strategy are: Extracting the potential sales value of each commodity bundle combination from the bundle sales optimization scheme to form a bundle commodity value data set; Based on the price sensitivity data of each consumer group and the bundled product value data set, a price elasticity analysis is performed to calculate the price adjustment point of each bundled product combination. The calculation formula is: ； in, represents the adjusted price, As the basic starting price, For consumer groups The change in demand, is the price elasticity of the consumer group, is the total number of consumer groups; According to the adjusted price, the sales price of each bundled product is adjusted to generate a price adjustment strategy.

8. The method for automatically pushing orders based on system connection based on big data according to claim 1, characterized in that: The steps for obtaining the personalized preferential recommendation results are as follows: Based on the price adjustment strategy, the discount content is customized according to each consumer's purchase record, behavior pattern and corresponding price sensitivity to form personalized discount recommendations; Implement personalized offer recommendations to consumer groups via email, social media, and in-app notifications; Collect and analyze consumers' acceptance rate of the personalized discount recommendations, determine the effectiveness of the discount implementation, and generate personalized discount recommendation results.

Images