원문정보
초록
영어
A lot of time of the users is consumed in searching appropriate papers related to the desired topic. It takes time to look through the paper also. In this paper, a hybrid method is introduced to classify research papers. This algorithm is designed to classify all research papers at the time of uploading in the repository. Hence it becomes easy to explore appropriate paper on a specific topic in minimum time. A data set has generated with research papers on different topics like natural language processing, machine learning, etc. The proposed algorithm passes the most frequent items fetched from the training data set to k-nearest neighbor method instead of the whole data set, to make clusters. The performance of the proposed method is compared with traditional KNN method which results the accuracy, improved by the factor of 7.46%.
목차
1. Introduction
2. Related Work
3. Problem Outline
4. Contribution
4.1. Prepare Dataset
4.2. Preprocessing
4.3. Mining Frequent Term Set
4.4. Create Document Term Matrix
4.5. Convert DTM To Data Frame
4.6. Apply KNN Algorithm
5. Application of the Proposed Algorithm
6. Experimental Results
7. Conclusion
References