@article{Aloysius_Tamilselvan_2022, title={A Novel Method to Reduce False Positives and Negatives in Sentiment Analysis}, volume={10}, url={https://ijisae.org/index.php/IJISAE/article/view/2177}, abstractNote={<p>Sentiment analysis focuses on the prediction of sentiment of the data by text processing, feature extraction, vectorization and classification techniques. The research areas in Sentiment analysis used to focus on the model that gives more accurate positive prediction. Reduction of False Positives and Negatives which are called Type I and II errors are not much dealt with. The best model may not be the best in reducing false positives and negatives. In this work rule-based and machine-learning algorithms are experimented to create a suitable model that gives equal importance to the reduction of positive and negative false values and accuracy. Experimented studies revealed that the model suggested using Linear Regression classifier that gave an overall accuracy of 62.35 % is the one that gives a better reduction in Type I and II error along with a competitive accuracy than SVM based model which was having the highest accuracy of 62.4%.</p>}, number={3}, journal={International Journal of Intelligent Systems and Applications in Engineering}, author={Aloysius, Chris and Tamilselvan, P}, year={2022}, month={Oct.}, pages={365–373} }