Design And Implementation Of Academic Recommendation System Based On Django Framework

At present,with the continuous development of society,mankind has entered an era dominated by big data.Massive academic data(papers,patents,journal conferences,etc.)appear on the Internet.It is more challenging than ever to establish scientific and effective cooperation in the computer field.It is usually difficult for researchers to find their best matching collaborators(BMCs).In addition,it is difficult for researchers to choose the right publishing place to publish their papers when they submit their papers,and it is a thorny problem to find the literature of interest in the massive papers.At present,most researchers search for useful academic resources by keyword search,but this method is inefficient.Therefore,there is an urgent need for an academic recommendation system to recommend interested academic resources for researchers to speed up the progress of scientific research.In this paper,we innovatively propose a best matching collaborators recommendation model based on multi-similarity fusion,which is called BMCRec,which integrates three academic characteristics: the similarity of cooperation relationship between scholars,the similarity of research field and the similarity of academic level.The similarity of cooperation relationship comprehensively considers two factors: the recent cooperation time between two scholars and the number of co-authored papers,and the similarity of research field The academic level similarity is measured by considering the content similarity of the abstracts published by scholars.The academic level similarity is measured by integrating five academic indicators,including the academic age of scholars,H-index,the number of published papers,the total number of citations and the number of collaborators.After that,the academic cooperation network is established through academic big data,and the transition probability matrix among scholars is constructed.Finally,the similarity between nodes in the network is calculated through the improved restart random walk algorithm,and compared with MVCWalker model and restart random walk model RWR,which proves the superiority of BMCRec algorithm.Finally,the recommended results are more accurate So as to recommend the best matching collaborators for the target scholars.Secondly,based on multi-layer perceptron,this paper establishes a journal conference recommendation model called JCR-MLP.JCR-MLP gives the title,abstract and keyword information of the paper to be submitted,and then extracts the feature information of the input text to form a feature vector by using natural language processing technologies such as TF-IDF,chi square test and hot coding.Finally,we input the feature vector into a multi-layer perceptron for prediction,and recommend appropriate journals or conferences to users for reference when submitting articles.Finally,based on the DBLP citation network data set,this paper implements the proposed algorithm of collaborator recommendation and journal conference recommendation,and designs and develops each function module of the academic recommendation system.The system can greatly improve the scientific research efficiency of scientific researchers,and has certain practical value.