Mobile O2O (Online to Offline) recommendation method and system

Abstract

The present invention discloses a mobile O2O (Online to Offline) recommendation method. The method comprises: a server automatically generating a consumption matrix according to a consumption list of a consumer, generating push information, which may be needed by the consumer, by commercial information of a merchant at a frequent consuming time for consuming items of the same type from previous consumption matrixes of the consumer, and sending the push information to the consumer at an appropriate time; and through a GPS, when a frequent consuming location is in a range that can be perceived by a mobile terminal carried by the consumer, the server sending the push information to the consumer by means of the mobile terminal. The present invention further discloses a mobile O2O recommendation system for implementing the method. According to the mobile O2O recommendation method and system disclosed by the present invention, accurate recommendation can be achieved, existing common recommendation systems based on content are combined, and, on this basis, a position attribute, a mobile perceiving attribute and the like are added, so that integrative utilization of online and offline resources of merchants is facilitated, O2O application is accurately implemented, and user viscosity is greatly strengthened.

Description

technical field [0001] The invention relates to location awareness based on user intelligent terminals, combined with historical transaction data to realize accurate mobile O2O recommendation in e-commerce systems and offline physical stores. Background technique [0002] The current mainstream recommender systems are mainly divided into the following four categories: [0003] (1) Content-based recommendation: through the user's search keywords, online criteria, consumption records, etc., find the most matching information from the background data and recommend it to the user; although the recommendation is highly accurate and does not require learning, its The essence is still passive recommendation, unable to discover the diverse needs of users. [0004] (2) Collaborative filtering recommendation: first classify users, and then use the consumption and evaluation records of other members in the classification to recommend products to users. Generally divided into two type...

AI Technical Summary

Problems solved by technology

Although it can achieve diversified recommendations, there is a cold start problem, that is, when the user is a new user who has just joined, the recommendation cannot be realized because he has not joined other groups

[0007] (3) Knowledge-based recommendation: Use some rules or examples in a specific field to implement recommendations. Although it is not necessary to establish a user demand preference model, it is difficult to formulate reasonable reasoning rules in the field

Images