Open AccessText Mining for Multiclass Research Paper Categorization
Author(s) -
Anshul Saxena,
Anamika Anamika,
Bhaskar Pant,
Vikas Tripathi
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.b7240.129219
Subject(s) - computer science , categorization , tf–idf , classifier (uml) , artificial intelligence , multiclass classification , support vector machine , text categorization , training set , information retrieval , task (project management) , document classification , natural language processing , stop words , word (group theory) , machine learning , data mining , linguistics , philosophy , physics , management , quantum mechanics , term (time) , economics , preprocessor
A research paper is a rich source of academic and innovative writing on a particular topic, and they are unstructured in nature. Categorization of documents refers to classification of documents in classes that are predefined. It is arduous for a user to categories research paper in different domains: because extracting meaningful and relevant words from the research paper is a challenging task. For extracting important information we have used certain methods and classifiers. Methods like bag of words and tfidf is used for processing data. Prepossessing the data includes string tokenizing and stop-word removal. Then the processed data is classified using SVM classifier. For multiclass classification; since predefined classes are 4, therefore 1-v-r classifier is used. The system performance is 88% with 800 training and 200 testing documents. It is analyzed that the model performs better when the training data is more. The aim of this work is to categorize the documents and allocate set of predefined tag to them. It also evaluates the performance of the model by considering different percentages for training and testing sets of documents.