期刊名称：Lecture Notes in Engineering and Computer Science
印刷版ISSN：2078-0958
电子版ISSN：2078-0966
出版年度：2019
卷号：2239
页码：200-203
出版社：Newswood and International Association of Engineers
摘要：The amount of data currently residing on social
media is not sufficiently tapped and is certainly limitless as
millions of people are constantly posting one message or the
other to these public forums on the Internet. Twitter is one of
the largest social media network with over 320 million monthly
active users which has proven to be a fertile ground for
harvesting opinion from several people to influence decisionmaking
process within organizations and institutions. Based on
a thorough review of literature and past work in the area of
text mining and twitter sentiment analysis, a system was
developed which applied three different supervised machine
learning algorithms to a dataset curated by graduate students
at Stanford University in order to accurately classify tweets
into either positive or negative sentiment based on its content.
The result showed that Maximum Entropy has the highest
accuracy of 83.5% among the three algorithms.
Based on further analysis and research it was discovered
that the classifiers could be improved upon. Using this as a
basis, the authors then implemented a system that learns from
wrong classification as corrected by the users. This paper
presents the results from this research.
关键词：Machine Learning; Learning by Induction;;
Supervised Learning Algorithms; Twitter; Naïve Bayes;;
Support Vector Machine; Maximum entropy.
