Using Full-Text of Research Articles to Analyze Academic Impact of Algorithms

摘要

Top-10 algorithms in data mining voted by experts were widely used in various domains. How about the academic impact of these algorithms in a special domain, e.g. Natural Language Processing (NLP)? To answer this question, this paper uses full-text corpus of research articles published in ACL conference to explore influence of the Top-10 data mining algorithms in NLP domain. Academic influence of algorithms is analyzed according to three aspects: number of papers which mention algorithm, mention frequency, and mention location of algorithm. What's more, we find the most popular algorithm in a particular task via correlation coefficient between algorithm and task. This research offers a new way for evaluating influence of algorithms quantitatively. Results show that there are obvious differences of influences among algorithms. Specifically, impact of SVM algorithm is significantly higher than the other algorithms. Moreover, the most related task resolved by each algorithm is different.