Product recommendation method, device and equipment based on an improved BERT model, and medium

Abstract

The invention relates to the field of artificial intelligence, and provides a product recommendation method, device and equipment based on an improved BERT model and a medium, which can pre-process initial data to eliminate noise, enable the initial data to meet the requirements of model training on data, and train the BERT model by using sample data based on a target loss function. The method achieves the accurate prediction of the life stage through the combination of an improved BERT model and a classification model, further assists in the recommendation of products, improves the prediction coverage, solves a problem that the coverage of an existing prediction scheme is narrow, carries out the product recommendation according to the next life stage of the user, builds the cognition of the user for the products in advance, and even, the user can be guided in advance to start to purchase a higher-order product, and pre-announcement with the product is realized, so that more potential customers can be mined, and the relationship with the existing customers can be maintained. The invention further relates to a blockchain technology, and the trained model can be stored on a blockchain node.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a product recommendation method, device, equipment and medium based on an improved BERT model. Background technique [0002] In order to better serve old customers and attract new customers, it is usually necessary to recommend products for customers. [0003] In the existing technology, most of the people are classified through some common tags, such as age, gender, consumption level, etc., combined with the interests of customers, through clustering or similar recommendation methods, to recommend the most suitable products for customers to buy . [0004] However, for financial or insurance products that are different from ordinary commodities, such as wealth management and insurance, the above method is not applicable. Since commodities tend to be consumption habits and hobbies, and financial thinking and insurance needs can be cultivated and guided in...

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present invention provides a product recommendation method, device, equipment and medium based on the improved BERT model, which can combine the improved BERT model and classification model to realize accurate prediction of life stages, and further assist in product recommendation, improving the The coverage of prediction solves the problem of narrow coverage of existing prediction solutions. At the same time, it recommends products according to the user's upcoming life stage, builds the user's awareness of the product in advance, and even guides the user to start buying higher-level products in advance. Products, to achieve pre-advertisement of products in order to tap more potential customers and maintain relationships with existing customers

Images