Sentiment analysis (or opinion mining) is a natural language processing (NLP) technique used to determine whether data is positive, negative, or neutral. Sentiment analysis has become an important task in text classification problems as it helps retailers understand attitudes, opinions, and emotions expressed within an online comment to provide a better shopping experience and increase sales. In this project, which was my bachelor's degree final project, I detected the polarity of the reviews on an online shop (DigiKala) using SVM.
- Data preprocessing
- Removing punctuations and non-alphabetic characters
- Replacing Arabic letters with Persian ones
- Removing stop words
- Feature extraction using the bag of words vectorizer
- Classifying features using SVM