TREC 2007 Blog Track Experiments at Kobe University
Generate an AI Snapshot to get a quick, structured summary of this paper.
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This paper describes the approaches to the opinion retrieval and blog distillation tasks for the Blog Track, and attempts to detect and discard spam blogs (splogs) and non-English blogs to improve system performance.
Abstract
This paper describes our approaches to the opinion retrieval and blog distillation tasks for the Blog Track. For opinion retrieval we employ a two-stage framework consisting of keyword search and opinion classification, where customer reviews collected from Amazon.com are used for feature selection. For the blog distillation task we consider all the blog posts belonging to a blog in order to estimate the relevance of the blog at large. To accomplish this, we first identify relevant blogs for a given topic by keyword search and then examine all the posts for each identified blog. In addition, we attempt to detect and discard spam blogs (splogs) and non-English blogs to improve system performance. 1