Extraction of Aspects from Drug Reviews Using Probabilistic Aspect Mining Model

Pooja Gawande, Sandeep Gore

Abstract: Reviews of medication from patients are numerous on the internet. This review provides a brief overview of approaches to aspect mining as they relate to drug discovery. Many adverse drug reactions on chronic diseases are not discovered during limited pre-marketing clinical trials, they are only observed after long term post-marketing investigation of drug usage. The detection of adverse drug reactions, as early as possible, is an important topic of research for the pharmaceutical industry. Mining significant topics from short and noisy reviews is big challenge. In light of this, such problem is addressed by proposing the probabilistic aspect mining model (PAMM) for identifying the aspects/topics relating to class labels. Because of unique feature of PAMM it focuses on finding aspects relating to one class only rather than finding aspects for all classes simultaneously in each execution. Besides the aspects found also have the property that they are class distinguishing, that means they can be used to distinguish a class from other classes. It helps to reduce the chance of having aspects formed from mixing concepts of different classes; hence the identified aspects are easier to be interpreted by people.

Keywords: Drug review, opinion mining, aspect mining, text mining, topic modeling