Date of Award

5-1-2014

Degree Type

Thesis

University or Center

Clark Atlanta University(CAU)

Degree Name

M.S.

Department

Computer and Information Sciences

First Advisor

Dr. Peter Molnar

Second Advisor

Dr. Roy George

Third Advisor

Dr. Hsin-Chu Chen

Comments

This research study is based on analyzing text on Twitter to quantify the correlation between the segregated opinions posted by the users on multiple issues and the news mention in between those opinions. The results show various correlation percentages within the range of sentiment used by the clusters of segregated opinions towards different topics. The study participates in fitting the model of mining the opinions into using multiple algorithms, such as Apriori for finding the trending topics, Hierarchical clustering to describe the semantic relatedness between adjectives and the Expectation-Maximization algorithm to mine the hidden variables which result into different clusters. The framework shows how those algorithms are applied on the dataset collected from Twitter. Multiple experiments are conducted with different filtering categories to extract tweets with certain properties suitable for the analysis.

Signature Location_Supplemental file.pdf (45 kB)
Notice to Users, Transmittal and Statement of Understanding

Share

COinS