Term Extraction and Course Supervisor : Jin Guo Clustering - PowerPoint PPT Presentation

COMP 762 Term Extraction and Course Supervisor : Jin Guo Clustering Presented by: Shruti Bhanderi

Overview

Text Chunking STS subcontractors shall track the progress of development activities in the progress report.

Processing of Extracted Terms STS subcontractors the progress development activities the progress report Remove determiners, cardinal numbers and possessive pronouns. STS subcontractors progress report progress development activities Lemmatization STS subcontractor progress development activity progress report

Compute Similarities between Terms 1. Syntactic Similarity Measures 2. Semantic Similarity Measures

Syntactic Similarity Syntactic Measures Distance-based Token-based Corpus-based Levenstein Cosine SoftTFIDF

Semantic Similarity LIN LCH HSO JCN WUP LESK PATH RES ❖ Is-a (vertical) and has-a (horizontal) relation chains (WordNet).

Clustering

Clustering algorithms ❖ K means. ❖ Hierarchical Clustering ❖ Expectation Maximization

K means

Hierarchical Clustering Divisive Top down Agglomerative Bottom up

Expectation Maximization

Thank you!