E-commerce precision marketing intelligent system and method fused with AI personalized recommendation

Abstract

The invention discloses an electronic commerce precision marketing intelligent system and method fusing AI personalized recommendation. The system comprises a multi-source data acquisition module, a data preprocessing module, a user portrait construction module, a real-time behavior monitoring module, a recommendation algorithm module, a marketing scheme generation module, a marketing effect evaluation module and a feedback optimization module. According to the method, the comprehensive user portrait is constructed by integrating the multi-dimensional user data, and the mixed recommendation algorithm is applied, so that the user requirements can be deeply mined, accurate personalized recommendation is realized, and the matching degree between the recommendation result and the actual requirements of the user is improved; by adopting the real-time behavior monitoring module and the recommendation algorithm module, the recommendation strategy can be updated in time according to the real-time behavior of the user, so that the recommendation better fits the current shopping intention of the user; by adopting the marketing scheme generation module, diversified marketing schemes can be generated, an effective marketing effect evaluation and feedback optimization mechanism is combined, marketing activities can be continuously optimized, and the marketing efficiency and income of merchants are improved.

Description

Technical Field

[0001] The present invention relates to the field of e-commerce marketing technology, and in particular to an e-commerce precision marketing intelligent system and method integrating AI personalized recommendation. Background Art

[0002] E-commerce marketing is a type of online marketing, the process of achieving marketing goals through the internet. The internet is fast, efficient, and low-cost, with information resources shared online and zero barriers to entry. As a new medium, the internet offers one-on-one interaction, a breakthrough compared to traditional media's focus on large audiences. From a marketing perspective, online one-on-one communication between producers and consumers allows for understanding customer requirements, desires, and suggestions for improvement, transforming the industrial era's mass production and marketing practices into small-group or even individual marketing. This provides significant consumer satisfaction and aligns with modern marketing concepts. The basic definition of e-commerce is a new business model that enables widespread global commercial trade, conducted through the open internet environment, using a browser / server approach, allowing buyers and sellers to conduct various transactions face-to-face, enabling online shopping for consumers, online transactions between merchants, online electronic payments, and various other business activities and related integrated service activities. E-commerce marketing is a type of online marketing, the process of achieving marketing goals through the internet through the process of achieving a series of marketing steps.

[0003] The existing e-commerce marketing system still has some problems. It does not mine user behavior data deeply enough and often makes recommendations based only on users' purchase records and browsing history, ignoring multi-dimensional data such as users' comments, interests and preferences on social platforms, thus affecting the marketing effectiveness of e-commerce. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of the existing technology and propose an e-commerce precision marketing intelligent system and method that integrates AI personalized recommendations.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions: an e-commerce precision marketing intelligent system integrating AI personalized recommendation, including a multi-source data acquisition module, a data preprocessing module, a user portrait construction module, a real-time behavior monitoring module, a recommendation algorithm module, a marketing plan generation module, a marketing effect evaluation module, and a feedback optimization module;

[0006] Multi-source data collection module, used to collect user behavior data and basic information within the e-commerce platform, as well as user speech data and interest tag data on external social platforms;

[0007] The data preprocessing module is used to perform cleaning preprocessing, deduplication preprocessing and normalization preprocessing operations on the collected multi-source data;

[0008] The user portrait construction module uses machine learning algorithms based on pre-processed data to construct user portraits that characterize user characteristics from the perspectives of basic attributes, consumption behavior, interest preferences, and social influence.

[0009] Real-time behavior monitoring module, which is used to monitor users' operation behavior on the e-commerce platform in real time and transmit the newly generated data to the recommendation algorithm module;

[0010] The recommendation algorithm module uses a hybrid recommendation algorithm that combines AI deep learning with traditional machine learning. It includes collaborative filtering algorithm, content-based recommendation algorithm, and deep neural network recommendation algorithm. It dynamically adjusts the algorithm weights based on the data transmitted by the real-time behavior monitoring module and outputs personalized recommendation results.

[0011] The marketing plan generation module automatically generates a variety of marketing plans based on the recommendation results output by the recommendation algorithm module and the merchant's marketing goals;

[0012] The marketing effectiveness evaluation module evaluates marketing effectiveness from the perspectives of sales conversion, user engagement, and user feedback after the marketing plan is implemented.

[0013] The feedback optimization module uses the reinforcement learning algorithm to optimize the recommendation algorithm module and the marketing plan generation module based on the results of the marketing effect evaluation module.

[0014] As a further description of the above technical solution:

[0015] The multi-source data collection module realizes data collection and synchronization by connecting with the e-commerce platform database and the third-party social platform API interface, and the collected external social platform data includes but is not limited to user-posted content, like and comment records, and follow list information.

[0016] As a further description of the above technical solution:

[0017] The data preprocessing module uses natural language processing technology to perform word segmentation, part-of-speech tagging, and sentiment analysis on text data, and normalizes numerical data. Sentiment analysis uses a method based on a combination of dictionary and machine learning to determine users' emotional tendencies towards products or marketing activities.

[0018] As a further description of the above technical solution:

[0019] When constructing the consumption behavior dimension portrait, the user portrait construction module mines the user's consumption cycle pattern through the time series analysis algorithm and incorporates it into the user portrait features.

[0020] As a further description of the above technical solution:

[0021] The deep neural network recommendation algorithm in the recommendation algorithm module adopts a self-attention mechanism model to weight the key information in user portrait data and product feature data to improve the accuracy of recommendations.

[0022] As a further description of the above technical solution:

[0023] The marketing plan generated by the marketing plan generation module also includes a personalized coupon issuance strategy. The personalized coupon issuance strategy predicts the user's response probability to different types of coupons based on the user's historical coupon usage data, and then dynamically adjusts the timing and type of coupon issuance.

[0024] As a further description of the above technical solution:

[0025] When evaluating user feedback indicators, the marketing effectiveness evaluation module extracts key problem categories from user reviews through text mining technology and quantifies and scores different problem categories.

[0026] An e-commerce precision marketing method integrating AI personalized recommendation, using the above-mentioned e-commerce precision marketing intelligent system integrating AI personalized recommendation, includes the following steps:

[0027] S1: Data collection and preprocessing: The multi-source data collection module collects multi-dimensional data from users, and the data preprocessing module performs cleaning, deduplication, and normalization preprocessing operations on the collected data;

[0028] S2: User portrait construction: The user portrait construction module uses machine learning algorithms to construct user portraits based on preprocessed data;

[0029] S3: Real-time behavior monitoring and recommendation: The real-time behavior monitoring module monitors user behavior data on the e-commerce platform in real time and transmits it to the recommendation algorithm module. The recommendation algorithm module uses a combination of multiple algorithms combined with real-time behavior data to dynamically adjust the recommendation strategy and output personalized recommendation results.

[0030] S4: Marketing plan generation and implementation: The marketing plan generation module automatically generates a marketing plan based on the recommendation results and the merchant's marketing goals and pushes it to the user for implementation;

[0031] S5: Marketing effectiveness evaluation: After the marketing plan is implemented, the marketing effectiveness evaluation module evaluates the marketing effectiveness from multiple dimensions;

[0032] S6: Feedback optimization: The feedback optimization module uses the reinforcement learning algorithm to optimize the recommendation algorithm module and the marketing plan generation module based on the marketing effect evaluation results.

[0033] The present invention has the following beneficial effects:

[0034] 1. Compared with the existing technology, the present invention builds a comprehensive user portrait by integrating multi-dimensional user data and uses a hybrid recommendation algorithm to deeply explore user needs, realize accurate personalized recommendations, and improve the matching degree between recommendation results and actual user needs.

[0035] 2. Compared with the existing technology, the present invention adopts a real-time behavior monitoring module and a recommendation algorithm module to timely update the recommendation strategy according to the user's real-time behavior, so that the recommendation is more in line with the user's current shopping intention.

[0036] 3. Compared with the existing technology, the present invention can generate diversified marketing plans by adopting a marketing plan generation module, and combined with an effective marketing effect evaluation and feedback optimization mechanism, it can continuously optimize marketing activities, improve the marketing efficiency and revenue of merchants, and at the same time enhance the shopping experience of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a framework diagram of the present invention. DETAILED DESCRIPTION

[0038] Reference Figure 1 The e-commerce precision marketing intelligent system integrating AI personalized recommendation provided by the present invention includes a multi-source data acquisition module, a data preprocessing module, a user portrait construction module, a real-time behavior monitoring module, a recommendation algorithm module, a marketing plan generation module, a marketing effect evaluation module and a feedback optimization module;

[0039] Multi-source data collection module, used to collect user behavior data and basic information within the e-commerce platform, as well as user speech data and interest tag data on external social platforms;

[0040] Specifically, it is used to collect user behavior data on e-commerce platforms, including but not limited to browsing history, purchase history, collection history, add-to-cart history, search keywords, evaluation content, user basic information such as age, gender, region, occupation, and user speech data and interest tag data on external social platforms such as Weibo, Douyin, Xiaohongshu, etc.; and realize data collection and synchronization through docking with the e-commerce platform database and the third-party social platform API interface, and the collected external social platform data is not limited to user posted content, like and comment records, and follow list information.

[0041] The data preprocessing module is used to perform cleaning, deduplication and normalization preprocessing operations on the collected multi-source data; the data preprocessing module uses natural language processing technology to perform word segmentation, part-of-speech tagging, and sentiment analysis on text data, and normalizes numerical data. Sentiment analysis uses a method based on a combination of dictionaries and machine learning to determine users' emotional tendencies towards products or marketing activities.

[0042] Specifically, the collected multi-source data undergoes pre-processing operations such as cleaning, deduplication, and normalization. For example, for text data such as reviews and social media comments, natural language processing techniques are used for word segmentation, part-of-speech tagging, and sentiment analysis. Numerical data such as age and consumption amount are normalized to facilitate subsequent data mining and analysis.

[0043] The user portrait construction module, based on pre-processed data, uses machine learning algorithms to construct user portraits that characterize user characteristics from the dimensions of basic attributes, consumption behavior, interest preferences, and social influence; and uses time series analysis algorithms to explore the user's consumption cycle patterns and incorporate them into user portrait features.

[0044] Specifically, the basic attribute dimension integrates basic information such as user age, gender, region, and occupation; the consumption behavior dimension analyzes the user's purchase frequency, average order value, consumption time period, preferred categories and other consumption behavior characteristics; the interest preference dimension combines the user's comments on social platforms and browsing and collection records on e-commerce platforms to extract the user's interest keywords and preferred product types; the social influence dimension analyzes the number of users' fans on social platforms, interaction frequency, and social relationship network to evaluate the user's social influence.

[0045] The real-time behavior monitoring module monitors user behavior on the e-commerce platform in real time and transmits newly generated data to the recommendation algorithm module, such as the currently viewed page, dwell time, and click behavior. Once new user behavior data is detected, it is immediately transmitted to the recommendation algorithm module so that the recommendation strategy can be adjusted in a timely manner.

[0046] The recommendation algorithm module adopts a hybrid recommendation algorithm that combines AI deep learning with traditional machine learning, including collaborative filtering algorithm, content-based recommendation algorithm and deep neural network recommendation algorithm. It dynamically adjusts the algorithm weight according to the data transmitted by the real-time behavior monitoring module and outputs personalized recommendation results; the deep neural network recommendation algorithm adopts the self-attention mechanism model to weight the key information in user portrait data and product feature data to improve the accuracy of recommendations.

[0047] Specifically, the collaborative filtering algorithm calculates the similarity between users based on the user's historical behavior data and recommends products that similar users are interested in to the target user; the content-based recommendation algorithm recommends products with similar content to users based on product attributes such as brand, material, function, etc. and the interest preferences of the user profile; the deep neural network recommendation algorithm builds a multi-layer neural network model, takes user portrait data and product feature data as input, learns the potential relationship between users and products through training, and outputs personalized recommendation results.

[0048] The marketing plan generation module automatically generates a variety of marketing plans based on the recommendation results output by the recommendation algorithm module and the merchant's marketing goals. It also includes a personalized coupon issuance strategy, which predicts the probability of users responding to different types of coupons based on their historical coupon usage data, and then dynamically adjusts the timing and type of coupon issuance.

[0049] Specifically, based on the recommendation results output by the recommendation algorithm module and combined with the merchant's marketing goals, such as new product promotion, clearance sales, and increasing average order value, a variety of marketing plans are automatically generated. Personalized coupon distribution strategies include setting coupons of different denominations and usage thresholds for different user groups, recommending product combinations such as combo packages, related product recommendations, and limited-time discounts.

[0050] The marketing effectiveness evaluation module evaluates the marketing effectiveness from the dimensions of sales conversion, user participation, and user feedback after the marketing plan is implemented. When evaluating user feedback indicators, it extracts key issue categories from user reviews through text mining technology and quantifies and scores different issue categories.

[0051] Specifically, sales conversion indicators: statistics on the number of orders, sales, conversion rate, etc. brought by marketing activities; user participation indicators: analysis of users' coupon collection rate, usage rate, click-through rate and add-to-cart rate of product pages, etc.; user feedback indicators: collection of users' satisfaction and opinions on marketing activities through user reviews, customer service feedback and other channels.

[0052] The feedback optimization module uses reinforcement learning algorithms to optimize the recommendation algorithm module and marketing plan generation module based on the results of the marketing effectiveness evaluation module. For example, if a certain marketing plan is found to have a low conversion rate, the relevant algorithm parameters and marketing plan strategy will be adjusted to improve the effectiveness of subsequent marketing activities.

[0053] An e-commerce precision marketing method integrating AI personalized recommendation, using the above-mentioned e-commerce precision marketing intelligent system integrating AI personalized recommendation, includes the following steps:

[0054] S1: Data collection and preprocessing: The multi-source data collection module collects multi-dimensional data from users, and the data preprocessing module performs cleaning, deduplication, and normalization preprocessing operations on the collected data;

[0055] S2: User portrait construction: The user portrait construction module uses machine learning algorithms to construct user portraits based on preprocessed data;

[0056] S3: Real-time behavior monitoring and recommendation: The real-time behavior monitoring module monitors user behavior data on the e-commerce platform in real time and transmits it to the recommendation algorithm module. The recommendation algorithm module uses a combination of multiple algorithms combined with real-time behavior data to dynamically adjust the recommendation strategy and output personalized recommendation results.

[0057] S4: Marketing plan generation and implementation: The marketing plan generation module automatically generates a marketing plan based on the recommendation results and the merchant's marketing goals and pushes it to the user for implementation;

[0058] S5: Marketing effectiveness evaluation: After the marketing plan is implemented, the marketing effectiveness evaluation module evaluates the marketing effectiveness from multiple dimensions;

[0059] S6: Feedback optimization: The feedback optimization module uses the reinforcement learning algorithm to optimize the recommendation algorithm module and the marketing plan generation module based on the marketing effect evaluation results.

[0060] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent e-commerce precision marketing system integrating AI personalized recommendations, featuring: It includes multi-source data acquisition module, data preprocessing module, user portrait construction module, real-time behavior monitoring module, recommendation algorithm module, marketing plan generation module, marketing effect evaluation module and feedback optimization module; Multi-source data collection module, used to collect user behavior data and basic information within the e-commerce platform, as well as user speech data and interest tag data on external social platforms; The data preprocessing module is used to perform cleaning preprocessing, deduplication preprocessing and normalization preprocessing operations on the collected multi-source data; The user portrait construction module uses machine learning algorithms based on pre-processed data to construct user portraits that characterize user characteristics from the perspectives of basic attributes, consumption behavior, interest preferences, and social influence. Real-time behavior monitoring module, which is used to monitor users' operation behavior on the e-commerce platform in real time and transmit the newly generated data to the recommendation algorithm module; The recommendation algorithm module uses a hybrid recommendation algorithm that combines AI deep learning with traditional machine learning. It includes collaborative filtering algorithm, content-based recommendation algorithm, and deep neural network recommendation algorithm. It dynamically adjusts the algorithm weights based on the data transmitted by the real-time behavior monitoring module and outputs personalized recommendation results. The marketing plan generation module automatically generates a variety of marketing plans based on the recommendation results output by the recommendation algorithm module and the merchant's marketing goals; The marketing effectiveness evaluation module evaluates marketing effectiveness from the perspectives of sales conversion, user engagement, and user feedback after the marketing plan is implemented. The feedback optimization module uses the reinforcement learning algorithm to optimize the recommendation algorithm module and the marketing plan generation module based on the results of the marketing effect evaluation module.

2. The e-commerce precision marketing intelligent system integrating AI personalized recommendation according to claim 1 is characterized by: The multi-source data collection module realizes data collection and synchronization by connecting with the e-commerce platform database and the third-party social platform API interface, and the collected external social platform data includes but is not limited to user-posted content, like and comment records, and follow list information.

3. The e-commerce precision marketing intelligent system integrating AI personalized recommendation according to claim 1 is characterized by: The data preprocessing module uses natural language processing technology to perform word segmentation, part-of-speech tagging, and sentiment analysis on text data, and normalizes numerical data. Sentiment analysis uses a method based on a combination of dictionary and machine learning to determine users' emotional tendencies towards products or marketing activities.

4. The e-commerce precision marketing intelligent system integrating AI personalized recommendation according to claim 1 is characterized by: When constructing the consumption behavior dimension portrait, the user portrait construction module mines the user's consumption cycle pattern through the time series analysis algorithm and incorporates it into the user portrait features.

5. The e-commerce precision marketing intelligent system integrating AI personalized recommendation according to claim 1 is characterized by: The deep neural network recommendation algorithm in the recommendation algorithm module adopts a self-attention mechanism model to weight the key information in user portrait data and product feature data to improve the accuracy of recommendations.

6. The e-commerce precision marketing intelligent system integrating AI personalized recommendation according to claim 1 is characterized by: The marketing plan generated by the marketing plan generation module also includes a personalized coupon issuance strategy. The personalized coupon issuance strategy predicts the user's response probability to different types of coupons based on the user's historical coupon usage data, and then dynamically adjusts the timing and type of coupon issuance.

7. The e-commerce precision marketing intelligent system integrating AI personalized recommendation according to claim 1 is characterized by: When evaluating user feedback indicators, the marketing effectiveness evaluation module extracts key problem categories from user reviews through text mining technology and quantifies and scores different problem categories.

8. An e-commerce precision marketing method integrating AI personalized recommendations, characterized by: An e-commerce precision marketing intelligent system integrating AI personalized recommendation according to any one of claims 1 to 7 comprises the following steps: S1: Data collection and preprocessing: The multi-source data collection module collects multi-dimensional data from users, and the data preprocessing module performs cleaning, deduplication, and normalization preprocessing operations on the collected data; S2: User portrait construction: The user portrait construction module uses machine learning algorithms to construct user portraits based on preprocessed data; S3: Real-time behavior monitoring and recommendation: The real-time behavior monitoring module monitors user behavior data on the e-commerce platform in real time and transmits it to the recommendation algorithm module. The recommendation algorithm module uses a combination of multiple algorithms combined with real-time behavior data to dynamically adjust the recommendation strategy and output personalized recommendation results. S4: Marketing plan generation and implementation: The marketing plan generation module automatically generates a marketing plan based on the recommendation results and the merchant's marketing goals and pushes it to the user for implementation; S5: Marketing effectiveness evaluation: After the marketing plan is implemented, the marketing effectiveness evaluation module evaluates the marketing effectiveness from multiple dimensions; S6: Feedback optimization: The feedback optimization module uses the reinforcement learning algorithm to optimize the recommendation algorithm module and the marketing plan generation module based on the marketing effect evaluation results.

Images