Detecting Types of News Using Hierarchical Machine Learning Model with Text Classification

  Subhadeep Chakraborty

News is one of the important aspects of human life from which they can gather the required information. In the early time, the news have been gathered by readers from the newspaper or from the news channels. With the advancement of technology, Social Media comes into the scenario from where readers can get their required news and many more things. In all cases, News can be of different types which are somehow difficult for the reader to identify. Machine Learning plays an important role here to detect the type of news with the implication of Natural Language Processing to analyze the text and to identify the types. In this research, the type of news has been detected by collecting the Insorts news database from Kaggle. The news texts have been prepared by cleaning and vectorizing with the implication of Term Frequency Inverse Document Frequency and Count Vectorization and the models of machine learning have been applied. In this context, the Hierarchical Machine Learning model has been proposed that combines the selected state-of-the-art models through Stacking Classifiers and Voting Classifiers. With the application of all state-of-the-art models and the proposed model, it has been observed that the proposed model has detected the type of news with the highest accuracy (94.22% using TFIDF and Unigram) which is also seen to be higher compared to the existing approaches

Keywords- Artificial Intelligence, Machine Learning, Text Analytics, Classification, Feature Extraction, Hierarchical Model, Model Overfitting

Unique Identification Number - IJEDR2204003
Page Number(s) - 20-31
Pubished in - Volume 10 | Issue 4 | November 2022
DOI (Digital Object Identifier) -    http://doi.one/10.1729/Journal.31970
Publisher - IJEDR (ISSN - 2321-9939)

  Subhadeep Chakraborty,   "Detecting Types of News Using Hierarchical Machine Learning Model with Text Classification", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.10, Issue 4, pp.20-31, November 2022, Available at :http://www.ijedr.org/papers/IJEDR2204003.pdf