EasyChoose: A Continuous Feature Extraction and Review Highlighting Scheme on Hadoop YARN

Today the Internet offers a massive amount of reviews and user experiences about a variety of products from different manufacturers, ranging from smartphones, automobiles, and home appliances to Internet services such as hotel booking and airplane booking. For a careful customer it is time-consuming to make good purchasing decisions due to a variety of similar products, lots of reviews for each product, and distributed reviews on the Internet. To alleviate this situation, this paper proposes EasyChoose, which is a distributed scheme based on Hadoop YARN to continuously collect product reviews from the Internet, extract representative product features based on previous customers' reviews, and highlight the main point of the reviews. In this paper, we use online hotel booking as an example to demonstrate the effectiveness of EasyChoose. The results show that EasyChoose is able to automatically extract representative product features and highlight reviews without losing the original meanings. Furthermore, EasyChoose is able to continuously provide such service to keep up with changes in recent customers' reviews.