Research On Personalized Learning Content Recommendation Based On Personality Traits

With the popularization of education informatization in recent years,learners’ learning methods have gradually shifted from offline classes to online learning,and due to the massive learning resources on the Internet,learners are prone to the information overload problem in the learning process.For online learners,the information overload problem will interfere with learners’ accurate selection of the learning resources they need due to the large amount of redundant information,thus affecting the effectiveness of learners’ online learning.Therefore,a personalized learning recommendation system is created to recommend the required learning resources for learners and improve their learning efficiency.In the current online learning process,a large amount of psychological information is often overlooked,and the interaction information generated by learners and online learning recommendation systems is rich in psychological characteristics,such as personality traits.Combining psychological features with personalized learning recommendation systems to form a "psychology +education" online learning approach is an important research direction for future learning recommendation systems.The use of learners’ personality traits to set personalized parameters for them and recommend learning resources that better meet learners’ psychological expectations is what personalized learning recommendation systems lack at present.This paper takes learners and learning resources of online learning websites as examples,integrates rich and diverse online education data,completes learner modeling and learning object modeling by extracting learner characteristics,and designs learning recommendation algorithm models for online learning recommendation based on learner and learning object models.In this paper,we make full use of the rich psychological information of learners to make learning recommendations for them,so that the learning recommendation system can be more personalized and achieve the integration of online education resources and learning content recommendations.The work of this paper includes the following points:1.In this paper,we model the learners by capturing the psychological Big 5personality traits of the learners,firstly,we extract the multifaceted features of the learners for online learning,and then we perform data pre-processing,text word embedding and PANDORA model training to build a learner model with the Big 5personality traits as the core features and the rest of the features as the auxiliary features.2.For the modeling of learning objects,the multivariate features of online learning resources were extracted,a BERT text word embedding method based on the selfattention mechanism was designed by data analysis,and a self-coding/decoding device was designed for further extraction of learning object features,and finally the core features were mapped,and a learning object model with Big Five personality features as the core features and the rest features as the auxiliary features was constructed.The learning object model is constructed with the Big 5 personality features as the core features and the remaining features as the auxiliary features.3.In this paper,a Pan-BF-PMF algorithm model based on the Big Five personality features is designed for the learning content recommendation algorithm,which makes full use of the psychological Big Five personality features hidden in the learners and has been proven to have excellent performance in recommending learning content for learners.Based on the comprehensive features of this model,the Pan-BF-PMF+algorithm model is optimally designed and experimentally explored on real data sets and compared with the baseline model,and the experimental results demonstrate its superior performance for learning content recommendation.Finally,the impact of the driving parameters designed in this paper on the experimental results is explored.