خط مشی دسترسیدرباره ماپشتیبانی آنلاین
ثبت نامثبت نام
راهنماراهنما
فارسی
ورودورود
صفحه اصلیصفحه اصلی
جستجوی مدارک
تمام متن
منابع دیجیتالی
رکورد قبلیرکورد بعدی
Document Type:Latin Dissertation
Language of Document:English
Record Number:54894
Doc. No:TL24848
Call number:‭1481859‬
Main Entry:Robert Stamper
Title & Author:Sentiment Mining of online political forumsRobert Stamper
College:University of Louisville
Date:2008
Degree:M.S.
student score:2008
Page No:37
Abstract:It has always been important to understand what people are thinking. In today's society, the internet has become increasingly active as a way for users to actively engage in political debates and discuss their political views. Political discussion forums have become a popular place for politically minded individuals to express their views. While political blogs are another popular spot for users to express their political views, well established political forums can have thousands of members and provide a wealth of data to be used for analysis. Sentiment Mining is used today to gather people's opinions in many areas. By using Sentiment Mining on political forums, we could gain a better understanding of how people might vote in upcoming elections, their party affiliations, and other political leanings. Sentiment Mining will be used to analyze two online political forums to try and find similarities and differences in political opinions of the forum members. The algorithm then can be used on other political forums to predict the political opinions of their forums members as well. The paper will also analyze current political happenings such as how economic crises or war may sway sentiments, and which topics are most important amongst each of the parties' websites. We will then compare various words and phrases of opinions to see which ones had a greater impact on being able to determine the political leaning of the forum. We found that through Sentiment Mining of online political forums, sentiment differed in many areas between forums, while other sentiments were important among both forums. We found out through Sentiment Mining on Republican and Democratic forums, that the Republican leaning forum, "Hannity" contained a high frequency of descriptive terms on terrorism, privacy, and Christmas, while the Democratic Underground forum had high frequencies of terms such as education, health, and Iraq. These terms show us what sentiments are more important to posters from each political forum.
Subject:Communication and the arts; Social sciences; Applied sciences; Political science; Information science; Computer science; 0984:Computer science; 0615:Political science; 0723:Information science
Added Entry:University of Louisville