Mining with myriads of internet data

Abstract

Ever since the Internet was open to common people, there has been an explosive growth of content on the web. Wikipedia alone holds information about more than four million topics. While it is sufficient at the person-level, the huge amount of data available in the web can be combined together from different corners and be used in internet-scale applications that can recognize patterns about world. In-order for this to be made real, topics on the internet should be made easily distinguishable from one another and should not rely on keywords. Freebase aims to assign a unique ID to each topic on the web much like how a bar-code is used on retail products. Since Freebase is manually created, it often lacks complete information. This project aims at filling one of those gaps and move a step closer to a complete knowledge base.