The Design And Implementation Of Computer Science Academic Rankings System

With the development of computer science today, there appears a lot ofconferences and journals held for computer science. These conferences andjournals publish tens thousands of papers every year and guide thedevelopment of computer science. How to make rankings and evaluate theseconferences, journals, institutes and universities is a hot and importantproblem. Although there exist such academic information and rankingsystems, there still appear a lot of blemishes. The coverage of papers isincomplete. Some published papers may be missing. Authors and theirinformation are not normalized, the ranking result is much different from thecurrent state. The papers’ category is not clear and it is hard to make rankingfor sub domain.This paper introduces Computer Science Academic Rankings (CSAR)which aims at academic information extracting, mining and ranking. In thispaper we mainly present approaches for information extraction,normalization and ranking algorithms in CSAR. We first list some popularacademic ranking systems and investigate their advantages and disadvantages.Then we propose the main problems and approaches in Internet academicranking. We design the CSAR system with four main components:information collection, information normalization, academic networkbuilding and academic ranking. By internet semantic analysis, we extractinformation of conferences, journals and papers automatically with webcrawler. With global geography and academic database, we normalize themulti-source data and improve the data quantity. Also we do paper classification with machine learning algorithm. Finally we make academicnetwork and publish accurate academic ranking.