A group meeting place intelligent recommendation method and system based on a large language model

By generating individual and location profiles through large language models and spatial optimization techniques, and combining them with mobile cost optimization for group centers, we have achieved high efficiency, fairness, and improved satisfaction when recommending meeting places for groups, thus solving the problem of balancing preference and convenience in existing technologies.

CN122240946APending Publication Date: 2026-06-19BEIHANG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIHANG UNIV
Filing Date
2026-04-01
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing intelligent assessment solutions struggle to simultaneously optimize preference matching and travel convenience when recommending meeting places for groups. They neglect differences in participants' mobility and transportation conditions, potentially leading to recommendations that do not align with participants' wishes or impose excessive travel burdens, thus reducing group satisfaction.

Method used

By analyzing individual language data and location evaluation data using a large language model, individual and location profiles are generated. By combining individual action data and road network travel data, the group center is optimized. Multiple rounds of screening and chain reasoning are performed to generate structured recommendation results, including preference matching, travel convenience, and conflict analysis.

Benefits of technology

Ensure that the recommended results align with the interests of the user group, reduce resistance from individual users, and improve the implementation and satisfaction of group meetings.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122240946A_ABST
    Figure CN122240946A_ABST
Patent Text Reader

Abstract

This invention discloses an intelligent recommendation method and system for group meeting locations based on a large language model. The method includes: outputting individual and location profiles from the large language model and inputting them into a spatial optimization model to obtain a group center; setting a search radius from the group center to generate a candidate location set, along with the individual movement cost and average movement cost corresponding to each candidate location; classifying and semantically annotating the candidate locations with tags to obtain a structured ranking result, and further generating a structured recommendation result. Thus, individual and location profiles are generated based on natural language analysis, and the movement cost of individual users to various locations is introduced for reverse weighted iteration to filter and obtain a candidate set with high accessibility. Then, multi-round chain reasoning is performed using full-process data to select highly feasible recommendation options from the candidate set that align with the interests of the user group, ensuring that group meetings can be successfully arranged without strong resistance from individual users.
Need to check novelty before this filing date? Find Prior Art