VisTopK – Visualization of Top-K Information Retrieval Algorithms November 3, 2006Posted by shahan in eclipse, eclipse plugin, GEF.
This project is for a course credit and is supervised by Dr. Consens. I will provide a visualization of Top-K results using static and dynamic views of Threshold Algorithm (TA) traces. This involves the use of inverted indexes from an XML document collection and will explore the effect of various factors such as compression, encoding, and retrieval/indexing algorithms. The deliverable will be a plug-in in Eclipse, due to its extensibility and ease of integration with various components. Current thoughts of the plugin are a main view containing a stock-chart and a property view. The visualization will be a front end to the XML format trace data collected during a separate TA run. Static and dynamic views will be provided for snapshots and temporal views. The visualization will be based on the depiction in:
IO-Top-k: Index-access Optimized Top-k Query Processing. Holger Bast, Debapriyo Majumdar, Ralf Schenkel, Martin Theobald, Gerhard Weikum.
An interesting paper which provides a good basis for TAs, the algorithms themselves are relatively easy to understand but requires acquiring the terminology of the context. The relation between, advantages, and differences of several different TA methods are discussed, some of which are No Random Access (NRA), Random Access (RA), and a combination of both.