Motivations

The Internet has witnessed the emergence of Web 2.0 social network sites, such as Facebook and Yelp. They allow individuals to construct a public profile and articulate a list of connected users to traverse and share contents within the system. They become very successful as they bring together social experiences from across small and disconnected off-line social networks. One the other hand, many mobile phones also generate mobile data during its interaction with environment, people simply use them unconsciously to do everyday tasks and leaving record of experience or problem about both people and the environment. For example, a customer can comment on a meal of a restaurant on Yelp.com with his mobile phone, which can be viewed by other Yelp users as a reference or recommendation. In general it is difficult for users to judge whether the information is useful to them or not in this huge information surge. People have to iteratively do mobile web search, whenever they need a piece of information.

To make people’s life better, we believe that there is a need to proactively extract information that is of interest to users. The project explores design of the framework BetterLife 2.0 that implements large scale social intelligence application on the Cloud environment. We adopted the Case-based Reasoning framework to providing logical reasoning. We outlined specific design considerations when porting a typical CBR framework on Cloud using Hadoop’s various services (e.g., MapReduce, HBases). These services allow efficient case base management (e.g. case insertion and adaptation) and computing intensive jobs distribution. With the scalability merit of MapReduce, we are able to provide recommendation service with social network analyzing for applications that can handle millions of users’ social activities.

Case Based Reasoning:

Case-based Reasoning (CBR) technique has been used in context-aware recommendation systems as a reasoning technique, which is the process of solving new problems based on the solutions of similar past problems. Similarity-based retrieval is a beneficial feature of case-based recommenders as it is suitable for problems where earlier cases are available, even when the domain is not understood well enough. With CBR, people would benefit from a contrast-and-compare analysis by supplying a previous case and its solution to convince a user to make a decision.