Classification And System Design Of Academic Report Notification

With the continuous popularization and development of science and technology in China,academic activities become more frequent.There are many websites of scientific research institutions,colleges,and universities,where academic report notifications will be updated every day.This dissertation mainly focuses on the classification of academic report notifications in order to let the researchers be able to get access to academic report notifications.The main work of this dissertation:(1)This dissertation first introduces background of the topic and theoretical basis of academic report notifications,and summarizes the classical algorithms of text classification process.(2)We classify and summarize the websites of academic report notifications,and design a special academic report web crawler for the distribution of elements inside the web page.(3)Thesis titles and abstracts from Wanfang Databases are used as training data to solve the problem of insufficient corpus.Furthermore,the influence of different model parameters,typical classification methods,and different corpus scales on the classification results are analyzed.(4)On the basis of the research work mentioned above,this dissertation designs a twopass classification algorithm to classify academic report notifications based on the classification confidence.Experiments show that our proposed method performs better than SVM and Naive Bayesian method respectively.(5)Finally,we build a website,including academic report crawler and classifier,which demonstrates the feasibility of our method.The system is displayed in the form of a website and a mobile client application.