My Active Research Projects

Ranked Query Processing

My active research towards my is on supporting ranked (i.e., top k) queries through the design of novel indexing algorithms and data structures. Computational Geometry forms the inspiration for the ideas, and the techniques are based on arrangements, duality, and range searching. Most recently, I have a paper at DASFAA which established some geometric foundations for using projection to evaluate the relevance of tuples to a user query.

Social Network Anonymisation

In collaboration primarily with Gautam Srivastava, I am studying social network anonymisation. Social networks are fast dominating the world wide web, which encourages substantial research. Releasing the data, however, induces privacy concerns, so effective measures to anonymise it are required. We have recently completed research on k -anonymisation of graphs that introduces an alternative means of distorting the original graph, which preserves a different set of structural properties. In addition to this, we have introduced a means of evaluating susceptibility to attribute disclosure attacks, alpha-proximity in a short paper to appear this year at ASONAM. We also make available the source code for the implementation in that paper.

Frequent Itemset Mining

My primary research during my M.Sc. was on modifying the A Priori frequent itemset mining algorithm to make it competitive with state-of-the-art implementations of alternative algorithms, such as FPGrowth, on large datasets. I have intentions of parallelising the work on a shared memory architecture, but this is on hold for now.