Text Classification Using Feature Extraction and Classification Model

Sentiment analysis represents context-based text mining that classifies and abstracts particular information from the source data, and helps to interpret people's feelings. It adopts the NLP approach to classify people's opinions regarding goods or feedback. Sentiment analysis handles people's opinion and behavior in the context of feelings and attitudes concerning an episode or incidence. There are many domains that use opinion mining widely such as commercial goods review, social media study and film evaluation and so on. Semantic analysis provides major support in building recommendation frameworks. User provides textual feedback such as web-based reviews, remarks or responses on social networking and business sites. Machine learning is a common approach used for classification. This project devises a nascent methodology through the integration of dissimilar frameworks. This work applies Random Forest as a feature extraction method for taking out crucial features of the data suite. The K-mean clustering is enforced which can cluster similar type of features and in the last voting classifier is applied. Voting classifier consists of LR, KNN, SVM models. Python is the tool used in this work for the implementation of the devised methodology and evaluates new methodology with respect to some global performance indices.