Presentation is loading. Please wait.

Presentation is loading. Please wait.

Disambiguation Algorithm for People Search on the Web Dmitri V. Kalashnikov, Sharad Mehrotra, Zhaoqi Chen, Rabia Nuray-Turan, Naveen Ashish For questions.

Similar presentations


Presentation on theme: "Disambiguation Algorithm for People Search on the Web Dmitri V. Kalashnikov, Sharad Mehrotra, Zhaoqi Chen, Rabia Nuray-Turan, Naveen Ashish For questions."— Presentation transcript:

1 Disambiguation Algorithm for People Search on the Web Dmitri V. Kalashnikov, Sharad Mehrotra, Zhaoqi Chen, Rabia Nuray-Turan, Naveen Ashish For questions visit: http://www.ics.uci.edu/~dvk Computer Science Department University of California, Irvine

2 2 Entity (People) Search Person1 Person2 Person3 Unknown beforehand Top-K Webpages

3 3 Standard Approach to Entity Resolution

4 4 Key Observation: More Info is Available =

5 5 RelDC Framework

6 6 Where is the Graph here? Use Extraction!

7 7 Overall Algorithm Overview 1. User Input. A user submits a query to the middleware via a web-based interface. 2. Web page Retrieval. The middleware queries a search engine’s API, gets top-K Web pages. 3. Preprocessing. The retrieved Web pages are preprocessed: a)TF/IDF. Preprocessing steps for computing TF/IDF are carried out. b)Ontology. Ontologies are used to enrich the Webpage content. c)Extraction. Named entities, and web related information is extracted from the Webpages. 4. Graph Creation. The Entity-Relationship Graph is generated 5. Enhanced TF/IDF. Ontology-enhanced TF/IDF values are computed 6. Clustering. Correlation clustering is applied 7. Cluster Processing. Each resulting cluster is then processed as follows: a)Sketches. A set of keywords that represent the web pages within a cluster is computed for each cluster. The goal is that the user should be able to find the person of interest by looking at the sketch. b)Cluster Ranking. All cluster are ranked by a choosing criteria to be presented in a certain order to the user c)Web page Ranking. Once the user hones in on a particular cluster, the Web pages in this cluster are presented in a certain order, computed on this step. 8. Visualization of Results. The results are presented to the user in the form of clusters (and their sketches) corresponding to namesakes and which can be explored further.

8 8 Correlation Clustering In CC, each pair of nodes (u,v) is labeled –with “+” or “-” edge –labeling is done according to a similarity function s(u,v) Similarity function s(u,v) –if s(u,v) believes u and v are similar, then label “+” –else label “-” –s(u,v) is typically trained from past data Clustering –looks at edges –tries to minimize disagreement –disagreement for element x placed in cluster C, is a number of “-” edges that connect x and other elements in C

9 9 Similarity Function Connection strength between u and v: –where c k – the number of u-v paths of type k –and w k – the weigh of u-v paths of type k Similarity s(u,v) is a combination

10 10 Training s(u,v) on pre-labeled data

11 11 Experiments: Quality of Disambiguation By Artiles, et al. in SIGIR’05 By Bekkerman & McCallum in WWW’05

12 12 Experiments: Effect on Search


Download ppt "Disambiguation Algorithm for People Search on the Web Dmitri V. Kalashnikov, Sharad Mehrotra, Zhaoqi Chen, Rabia Nuray-Turan, Naveen Ashish For questions."

Similar presentations


Ads by Google