| In order to solve the problems that the researchers encounter in the search of scientific papers,the personalized recommendation of scientific papers has become one of the hot researches.Aiming at the importance of mathematical expressions on the development of scientific and technological information as well as in order to improve the accuracy of the recommendation of scientific papers,a method that is combined with mathematical expressions characteristics based on the traditional recommendation is proposed.Firstly,the method makes use of the user' historical queries which are based on mathematical expressions,the formulae's characteristics are extracted to generate the feature vectors which will lay the foundation for the calculation of the similarity of the papers.Secondly,the method analyzes the user's downloaded papers and historical queries,and uses the ontology and feature vector to form a model of the user interest,then to find the user's similar neighbors.Finally,the recommended algorithm is designed and recommended the papers to the user who may be interested.At the end of this thesis,the evaluation index of the recommendation system—recall rate,accuracy rate and the F1_Measure index are used to measure the recommended method for verifying the method's effectiveness. |
PDF Full Text Request