An Analytical Study Of Feature Extraction Techniques For Student Sentiment Analysis

Abstract

Student satisfaction is an important factor in the Web-Based Learning System(WBLS). Hence feedback of the Students plays a vital role in the measurement of the effectiveness of any WBLS. The Analysis of feedback or comments is known as Sentiment Analysis (SA) or Opinion mining.SA is the application of NLP used to identify the opinion or emotions behind the comments. Sentiment analysis is a text classification tool that focuses on the polarity of the text (positive, negative, neutral ) also emotions (happy, sad, angry)., Classification can be binary (positive or negative) or multi-class. In This paper, we applied two types of Feature Extraction Technique (FETs) namely Count Vector (CV) or Bag of Word (BoW) and Term Frequency and  Inverse Document Frequency (TF-IDF).Also presented a comparative analysis of the performance of the machine learning algorithms like Logistic Regression (LR), Support Vector Machine (SVM), Naive Bayes(NB), Decision Tree (DT) over Web-based learning models to classify the Student Feedback Dataset (SFD), emphasis is given on the sentiments present in the feedback of the students.