Fine-grained topic detection in news search results
Jia Cheng, Jingyu Zhou, Shuang Qiu · 2012
Current news search engines return results in many clusters, but often not very accurate. This paper studies fine-grained topic detection within news search results, which faces the challenges of short result length and highly related content. We propose an agglomerative clustering algorithm with a novel combination of similarity measures and use simulated annealing for optimization. Experimental results demonstrate that our approach can significantly outperform original search engine results.