A0391
Title: Bayesian multitask learning for medicine recommendation based on online patient reviews
Authors: Yichen Cheng - Georgia State University (United States) [presenting]
Yusen Xia - Georgia State University (United States)
Xinlei Wang - Southern Methodist University (United States)
Abstract: A drug recommendation model is proposed that integrates information from both structured data (patient demographic information) and unstructured texts (patient reviews). It is based on multitask learning to predict review ratings of several satisfaction-related measures for a given medicine, where related tasks can be learned from each other for prediction. The learned models can then be applied to new patients for drug recommendation. This is fundamentally different from most recommender systems in e-commerce, which do not work well for new customers (referred to as the cold-start problem). To extract information from review texts, both topic modeling and sentiment analysis are employed. Variable selection is further incorporated into the model via Bayesian LASSO, which aims to filter out irrelevant features. To the best of knowledge, this is the first Bayesian multitask learning method for ordinal responses. Multitask learning is also applied to medicine recommendations for the first time.
