The emergence of social networks has led to a considerable expansion of biomedical textual information. A new entity has been detected within such expression, namely, adverse drug reactions (ADRs). ADR can be defined as the side effect that can be mentioned in people’s opinions. Extracting such entities along with categorising their polarities (positive or negative) can offer the opportunity to obtain valuable feedback on important drugs and medicines. In this book, a method for ADR extraction and sentiment classification is discussed. First, an extension of trigger terms as features. Second, an enhancement of the medical lexicon is applied using a pre-trained model of biomedical word embedding. Third, a new document embedding approach is proposed based on Recurrent Neural Network in order to improve the medical sentiment document vectors. Through this book, the new methods of doing ADR and medical sentiment analysis on social media text is shown to have helped in extracting ADR more efficiently and classifying the sentiment polarity more accurately. This book is useful as a reference for those working in mining opinion on medical text, specifically on patients’ opinions over social media. In addition, it provides an interesting read for natural language processing enthusiasts working on medical text since the ideas in this book are in line with the current hot and popular approaches in natural language processing: word embedding and deep learning.