Design And Implementation Of Two-way Recommendation System For Teachers And Students On A Hybrid Model

With the increasing number of graduate students,the issue of mutual selection between teachers and students in colleges and universities has become more and more important.At present,there is a relatively single way for students to obtain tutor information.They mainly use search engines to enter the college's official website to obtain tutor information.This method can meet the small-scale mutual selection needs of teachers and students,but there is information overload and retrieval efficiency in large-scale mutual selection scenarios.Low-level issues.Due to the unclear goals,many students adopt the strategy of broadly submitting resumes,resulting in teachers receiving a large number of resumes and increasing the cost of tutor screening.During the master's degree,with the goal of engineering practice,I worked hard to apply theoretical knowledge to practice,and independently designed and developed a two-way recommendation system for teachers and students based on a mixed model.This thesis applies personalized recommendation to the two-way selection of teachers and students,avoiding students' blind selection of tutors and reducing the cost of tutor screening.This system adopts the browser and server architecture to design,adopts the front-end and back-end separation method to develop,integrates the visualization of knowledge graph,and realizes a new way of presentation.The main modules of the system include user management module,information collection module,information visualization module,teacher-student two-way recommendation module,and page interaction module.This article puts forward two innovations around the two-way recommendation platform for teachers and students:1.In the information visualization module,the system introduces the concept of knowledge graphs,extracts teacher information,defines teacher names as entities,and the edges between entities represent the similarity relationships between teachers in the field,which is convenient for students to understand the fields of interest.The tutor group improves the retrieval efficiency.2.In the teacher-student two-way recommendation module,this paper uses the topic model matching method based on LDA to solve the user cold start problem in the traditional recommendation algorithm,and uses the collaborative filtering algorithm based on K-Means clustering to solve the problem of matrix sparseness,and weights time The coefficients are introduced into the hybrid weighted model to solve the problem of user preferences changing over time,and finally the harmonic averaging algorithm is introduced to improve the probability of two-way matching between teachers and students.Finally,this thesis tests the system.The test results show that the system can meet the basic needs between teachers and students,and the recommendation algorithm based on the hybrid model can achieve high-efficiency teacher and student recommendation.