Active learning in E-Commerce Merchant Classification using Website Information

Abstract

Data and the collection and analysis of data plays an important role in everyday life even though it often goes unseen. In our case, our partner is using data to classify websites into different categories. We used active learning and other machine learning methods to help classify websites into these categories and to explore the data collection and classification process. We scraped text data from websites, translated the data to English, and then worked with machine learning tools to understand the data and classify it. We found that the xPAL active learning strategy and linear support vector classifiers seemed to perform best with our data. 1