Work

Quantitative Encoding of Bags-Of-Words for Sentiment and Sarcasm Detection in Textual Data

Year: 2024
Type: dissertation
Abstract: Natural language processing is the field of computer science dedicated to the computerized understanding of textual data. In this thesis, we introduce a novel tool for natural language processing involving the Jaccard distance, a measure of similarity between sets, and multilateration, a generalization of the concept of trilateration of the plane to general metric spaces. By abstracting sentences as sets of words, we are able to create a relatively low-dimensional numerical representation of text with potential applications to sentiment analysis and sarcasm detection. This thesis has been partially funded by the NSF grant No. 1836914.
Author: Paradise, Alexander Joseph
Language: English
Open Access status: green
OpenAlex ID: W7113320812
Cites: 5
Cited by:
Related to: 10