ADVANCE Fellows Award

Stanford University, Stanford CA

Investigators

Abstract

0137761 Mishra, Mina Stanford University ADVANCE Fellows: Data Mining Algorithms for Business Intensive Applications This proposal, framing business intelligence applications as concrete theoretical problems, develops data mining techniques for applications in customer management, wireless mining, and product affinities. The work, studying clustering measures, alternate embedding techniques, and computational complexity, involves designing computationally efficient algorithms that find approximately good answers. The scalability and I/O efficiency of proposed algorithms will also be studied. Two problems are mentioned for: (1)Identifying wireless communities and product affinities: the Maximum Edge Bi-clique problem and (2) Customer segmentation, text clustering, new clustering measures that combine inter and intra distances: the Conjunctive Clustering problem. The algorithms can be applied to various areas, including e-commerce. Results from this work are expected to impact (3) Customer relationship management by segmenting the customer population to target marketing programs more effectively and (4)Cell phone usage data to enable wireless companies to discover patterns in wireless access.