Self-adaptive companion mutual assistance recommendation method based on learning portrait matching

By identifying and isolating students' short-term language anomalies, a learning profile based on stable expressions is generated. An adaptive matching mechanism is introduced to solve the problem of misjudgment of learning profiles and achieve the stability and self-optimization of peer-assisted recommendation.

CN121834052APending Publication Date: 2026-04-10阜阳幼儿师范高等专科学校
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-26
Publication Date
2026-04-10

AI Technical Summary

Technical Problem

In existing technologies, students' short-term abnormal language expression is not identified, leading to misjudgment of learning profiles and affecting the accuracy and stability of peer-assisted recommendations.

Method used

By establishing a continuous time sequence of emotional stability, abnormalities such as sudden changes in tone, abrupt reduction in vocabulary, and abrupt sentence structure are identified, and a list of emotional anomalies is generated. In the process of learning and profiling, the feature sampling range is limited, a shadow matching channel and a profiling self-adjustment link are introduced, and the expression tone record is used as the primary method for matching.

Benefits of technology

It effectively prevents short-term emotional fluctuations from interfering with the learning profile, ensures the adaptive adjustment and anti-interference ability of the recommendation process, and maintains the stability and consistency of peer-assisted recommendation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121834052A_ABST
    Figure CN121834052A_ABST
Patent Text Reader

Abstract

The invention discloses a self-adaptive companion mutual assistance recommendation method based on learning portrait matching, and relates to the technical field of educational informatization and intelligent recommendation, and the method comprises the following steps: building a time continuous sequence of text records generated by students in different time periods, dividing the sequence into a stationary stage and a disturbance stage according to the emotion stability degree, and combining the learning scene label to generate an emotion fluctuation trajectory, and forming an expression sequence with a time boundary. According to the method, the emotion avoidance track and the expression tone record are constructed to realize identification and isolation of short-term expression anomalies, so that the learning portrait is generated only based on stable expression, and the stability and authenticity of the portrait are ensured. Meanwhile, a shadow matching channel and a portrait self-adjusting link are introduced, the weight is dynamically adjusted when the matching result is different, the self-adaption and the interference immunity of the recommendation process are kept, and long-term balance and continuous optimization of the learning portrait are achieved.
Need to check novelty before this filing date? Find Prior Art