Date of Submission

Spring 2013

Academic Program

Computer Science

Project Advisor 1

Sven Anderson

Abstract/Artist's Statement

Similarity is nominally the quality or condition of two things bearing likeness to one another. Semantic similarity, then, is a term used to determine how alike they are in what ideas they convey. Using semantic graphs, it is possible to create a method in which one can calculate semantic similarity algorithmically. ConceptNet is a large, human-generated graph that provides semantic and contexual information about lingual items called concepts. ConceptNet is a semantic graph on which the similarity between sentences is calculated. The calculation is performed by first ranking the node components, each representing a concept, to filter them down to what is essentially a sentence's semantic kernel. Then measure the semantic similarity of node pairs by calculating the length of the pathway between them as well as their proximity to other important nodes, summing up each pair's score in order to obtain a semantic rating for the sentences. Both word-to-word and and sentence-to-sentence semantic similarity measurements were accurate enough to warrant further investigation

Distribution Options

Access restricted to On-Campus only

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

Download

This work is protected by a Creative Commons license. Any use not permitted under that license is prohibited.

Bard Off-campus Download

Bard College faculty, staff, and students can login from off-campus by clicking on the Off-campus Download button and entering their Bard username and password.

Share

COinS