SUVANAM SASIDHAR BABU

Department of Computer University of Gondar, Gondar Ethiopia

Spatiotemporal text mining, web logs, mixture model, theme pattern, Application Defined Query (ADQ)

December 2010

When drilling for data via SQL injection there are three classes of attack in band, out-of-band and the relatively unknown inference attack. In band attacks extract data over the same channel between the client and the web server, for example, results are embedded in a web page via a union select. Out-of-band attacks employ a different communications channel to drill for data by using database mail or HTTP functions for example. Inference attacks stand alone in the fact that no actual data is transferred rather, a difference in the way an application behaves can allow an attacker to infer the value of the data. SQL Inference is the subject of this paper; this paper is the paper I promised I'd write after talking about this at the Blackhat Security Briefings in Europe in the March of 2005. Better late than never. Mining subtopics from web logs and analyzing their spatial temporal patterns have applications in multiple domains. In this paper, we define the novel problem of mining spatiotemporal theme patterns from web logs and propose a novel probabilistic approach to model the subtopic themes and spatiotemporal theme patterns simultaneously. The proposed model discovers spatiotemporal theme patterns by (1) extracting common themes from web logs; (2) generating theme life cycles for each given location; and (3) generating theme snap-shots for each given time period. Data mining holds the key to uncovering and cataloging the authoritative links, traversal patterns, and semantic structures that will bring intelligence and direction to our Web interactions.

2455-9997 (Online) - 2229-3531 (Print)

Vol.1 - No.2

Click here to download full paper