User post comments on social media in the form of reviews or comments to share their opinions on any sectors. These comments are referred as sentiments/opinions and it has a huge impact in all business fields. Basically, the sentiments are of three types namely positive, negative and neutral or it may be fine grained as ratings. These sentiments should be analyzed to take decisions for further improvement in business. But, this type of data produce massive stream hence machine learning techniques are employed in finding the hidden interesting pattern behind this comments. User reviews are tagged with sentiments or scores using Natural Language Processing (NLP) and then it is classified by data mining supervised machine learning techniques. This research work focuses on hotel industry in which users post comments and views in online mode regarding the hotels they stayed. This work uses machine learning approach TF-IDF with bi-gram for word level sentiment analysis and classifies the sentiments using Naive bayes classifier combined with ensemble method. This classification model will aid in better decision support system. The research work uses Rapidminer tool for visualization, implementation, and evaluation.