COMPA: A Comparative Retrieval and Analytical Engine for Consumer Products

Ramses J. Sotelo Jimenez, Bhavya Medishetty, Byron J. Gao

Abstract: Often people need to make comparisons in order to make selections and decisions. Increasingly such comparison activities are conducted online. For example, consumers typically read many reviews and compare various models before an online purchase. Since this comparison-based analytical and decision-making process often involves significant manual effort, it is greatly beneficial if the process can be largely automated and effectively performed by computers. In this research, we make an effort to design and implement COMPA, a web-based comparative retrieval and analytical engine in the context of consumer products that can effectively assist decision-making. COMPA takes two products A vs B as input, retrieves relevant online reviews, performs comprehensive analysis, and returns a comparison report including product comparison scores as well as supporting evidence.