My interest in social network analysis started with an internship at HP labs, where we were working on the problem of incentivizing users for optimal query routing. Routing in social networks is different because a social influence (or a sense of trust) is associated with the receipt of a message. However to make the incentive mechanism budget feasible, requires finding a finite set of nodes which when incentivized can achieve the desired routing to a target subset of nodes. Our work on an incentive based multi-target routing system aims to address this problem, by finding the subset specific top-k influential and implementing on a real world network. In the recent times, social gaming has gained significant popularity; we present one such app using Facebook where users can predict the score of a live cricket match. Nowadays, almost all access to user's data is handled using the OAuth protocol, which we elicit in a sample Twitter application using Google app engine and Java programming language. Following is the list of projects in this page,
An Incentive Based Multi-Target Routing System for Social Networks
Predicket – Predict Live Cricket Scores With A Social Experience
An Incentive Based Multi-Target Routing System
For Social Networks
This work was done for HP Labs, Bangalore as part of my Master’s (MTech) at IIIT-B. I thank my mentor Mr. Praphul Chandra (HP Labs, India) for his guidance and contribution to the project.
Abstract:
Recent trends in usage of online social networks show a tremendous potential for targeted advertising. The goal in a targeted advertising campaign is to expose aspecific content or ad to a desired set of users identified based on for example, demographics information. A set of influential users are then seeded with the content such that a maximum spread of the content in the desired set of users is achieved. We address a generalized form of this problem where the aim is to find the top-k influential nodes which maximize the influence on a given sub-set of nodes in a network. We propose an iterative network pruning approach to find the subset specific top-k influentials. We compare our algorithm with other approaches adapted to the said problem on two real world data sets and show the use of a tunable parameter - 'gamma' to make a trade-off between efficiency and performance. We devise an analysis framework for our approach and use it to show that the influence spread function remains to be sub-modular even when the graph is being pruned. Finally, we give a software design and implementation of the overall system and briefly discuss about incentive mechanisms to reward intermediate nodes which assist in routing the content. Our work although specific to influential detection in social networks, can be useful in other domains like epidemiology, approximation algorithms, software engineering, etc.
Downloads: Masters Thesis Source Code
Publications:
P. Chandra and A. Kalyanasundaram, "A Network Pruning Based Approach for Subset Specific Influential Detection", in 4th Annual ACM conference on Web Science (WebSci 2012), Evanston, Illinois, USA, Jun. 2012. [PDF] [PPT]
Predicket – Predict Live Cricket Scores With A Social Experience
This is a fun game developed to learn information extraction from the web and online social networks. I observed that most people, especially at work prefer to view cricket scores online. To make it fun and exciting I came up with this interactive score prediction game – Predicket. The idea is simple, the user can predict the score of a match while it is in progress. They win points based on the accuracy of their prediction. Users can then share their predictions with their friends. Work is in progress to allow users to compete with their friends in real time.
Downloads: Source Code Monetization Proposal Screenshot(1) Screenshot(2)
Click Here to play this game now.
Simple Twitter Application using Java and Google App Engine
This is a sample project to build a Twitter application using Google App engine and the Java programming language.
Click Here to access this app.
Downloads: Source Code