Predicting Polarity of Social Media Data. Using Statistical Feature Optimization for Efficient Sentiment Analysis

      Аннотация: In the era of social web that includes social networks, forums and blogs, the sentiment analysis is critical in decision making activities by an individual or an organization Sentiment analysis is an information retrieval technique that delivers the vision of relevant users like customers regarding entities like products or services. With the phenomenal growth in the data quantity of social web, manually analysing the opinion is almost impractical. In this context the automated sentiment analysis become critical research objective that grabbed researcher’s attention over a decade. Further, with respect to this, the contribution aims to design learning approaches based on Machine Learning for explorative Twitter trends sentiment analysis. Many of the contemporary SA methods envisioned the intricacies because of maximum feature volume. This gap has been addressed by Feature Selection and optimization using statistical assessment schemes for choosing optimum features.

Ключевые слова: Sentiment Analysis, Machine Learning, Feature Optimization, Feature Selection