Text Categorization using Supervised Machine Learning Techniques
Ishaan Dawar, Narendra Kumar, Sakshi Negi, Sayeedakhanum Pathan, Shirshendu Layek
Abstract
Text categorization is a task for text mining that involves pattern classification and is essential for the effective management of textual information systems (TIS). Each document is automatically assigned one or more categories from a set of predetermined categories as part of TIS. There is increased interest in creating tools to aid people in more effectively finding, filtering, and managing existing digital resources as the amount of information is rising. Those developed tools are involved in the administrative sectors and social benefits. This paper provides a general overview of text categorization models using various supervised machine learning techniques, including Logistic Regression (LR), Naive Bayes (NB), Random Forest (RF), Support Vector Machine (SVM), and AdaBoost, and compares their performance on various metrics, including accuracy and precision. The comparative analysis shows that the most effective algorithm for text classification is NVM with the highest accuracy of 96.86%, furthermore, AdaBoost is the worst in this case study with its minimum accuracy of 74.49%. We have also shown a comparison of other supervised machine-learning algorithms in this paper.