Bilkent University
Department of Computer Engineering
S E M I N A R

 

Labeling Search Result Clusters with Data Fusion

 

Anıl Türel
MSc. Student
Computer Engineering Department
Bilkent University

Search result clustering is a way of presenting information retrieval results via clustering them. Aim is to enhance search experience of users and decreasing time needed for search. Clusters contain related documents and for each cluster, according to documents it contains, a representing label is needed, which gives information about the cluster. Users can diminish size of search results by finding the cluster they are looking for. Hence, labels of clusters have great importance for giving the correct information about the cluster. In our method, we use data fusion for labeling clusters. After data fusion of documents in the cluster, we obtain a fused document that contain important terms in a ranked order. Combining highest ranked terms of the fused document gives the label of the cluster.

 

DATE: 20 December, 2010, Monday @ 14:30
PLACE: EA409