Machine Learning Based Automatic Classification of Customer Sentiment

Rapid growth of the use of internet has generated the enhancement of customer satisfaction as a fundamental issue in the modern era. Customer reviews can provide resourceful information for monitoring as well as enhancing customer satisfaction levels and thus business organizations can improve customer and product services to their utmost level. Sentiment analysis can manipulate these customer reviews by extracting product aspects and identifying the sentiments as positive or negative and additionally, text mining is used to switch unstructured text into structured form for this purpose. The principal goal of this paper is to perform opinion classification. Bag of Words (BOW) and Term Frequency-Inverse Document Frequency (TF-IDF) feature extraction techniques are used along with N-gram and Support Vector Machine (SVM) is regarded for review classification. Customer reviews from different restaurants of Dhaka were collected to accomplish the task. We believe that the result from the experiment can point out the effectiveness of the implemented technique of customer review analysis by saving the time as well as minimizing the difficulties in measurement of customer satisfaction.