A place real-time recommendation method and system based on combination weighting and user feedback

By combining user feedback and dynamic weight adjustment, and utilizing eye-tracking data, user actions, and text input, a high-precision list of location recommendations is generated. This solves the problem of insufficient accuracy in recommendation results in existing recommendation systems and achieves personalized and real-time recommendation effects.

CN119312895BActive Publication Date: 2026-06-05NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF 8 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
Filing Date
2024-10-18
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

Existing recommendation systems lack the ability to dynamically adjust weights and provide personalized recommendations when processing diverse data, resulting in inaccurate recommendation results that fail to meet users' high demands for real-time performance and personalization.

Method used

A dynamic weight adjustment mechanism based on user behavior is adopted, combined with an adaptive learning algorithm based on user feedback. An initial location attention ranking list is generated through eye-tracking data, operation behavior, and text input. The subjective and objective weights of modal information are calculated using the G1 and CRITIC algorithms. The difference coefficient method is used for dynamic fusion, and the weights are updated based on user feedback to improve the accuracy of the recommendation system.

Benefits of technology

It significantly improves the relevance and accuracy of location recommendations, meets users' needs for real-time and personalized services, and generates information with high coverage and high precision.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119312895B_ABST
    Figure CN119312895B_ABST
Patent Text Reader

Abstract

The application discloses a place real-time recommendation method based on combination weighting and user feedback, which comprises the following steps: extracting features of place-related data of three modes of eye movement data, operation records and text data in a period of time, and obtaining place attention degree ranking lists under each mode; initializing subjective weights; calculating objective weights; comprehensively calculating subjective weights and objective weights by using a difference coefficient method, and calculating an output place recommendation list; obtaining feedback information of a user on the place recommendation list, calculating normalized loss cumulative gains of the feedback information and the place attention degree ranking lists under different modes, and updating relative importance degrees and subjective weights of three modes of information. The application can recognize, comprehensively process and fuse data from a multi-element acquisition system, and obtain more comprehensive, accurate and reliable information than any single data source, so that the inherent shortcomings of a single data source are effectively made up, and high-coverage and high-precision information is generated.
Need to check novelty before this filing date? Find Prior Art