Sentiment analysis and opinion mining has become a major tool for collecting information from customer reviews on user sentiments and emotions, especially for online video streaming services and social networks. The increasing use of smartphones has popularized subscription to various streaming services that provide streaming media and video-on-demand. These applications offer a gateway to analyze user reviews by introducing sentiment analysis in the mobile environment. Online user reviews can hold a lot of useful information and help predict user interests. Analysis of user reviews can provide substantive information for business processing. Sentiment classification of these reviews is a commonly used analysis technique. Usually, these reviews are given in a text format, with every word in each considered a feature, so selection should focus on optimal features from all available features present in the reviews. This study employs machine learning algorithms to extract the best features from the training review data set. Then, the selected features are fed into the convolutional neural network and other fully connected layers for further processing. The proposed approach is evaluated with the standard evaluation metrics, such as precision, accuracy, recall, and f-measure, using three distinct benchmark data sets: polarity, Rotten Tomatoes, and IMDb. This work has also employed a pretrained sentiment analysis model over an Android application framework to classify reviews on a Smartphone without the need for any cloud or server-side API. © 2019 John Wiley & Sons, Ltd.