Discovering a Way to Analyze Customer Emotions on Social Media for use in Advertising Systems
Subject Areas : Food Security
leila khajehvand
1
(
M.A.,Department of Management Information Systems, Science and Research Branch, Islamic Azad University, Tehran, Iran
)
Abbas Toloie Eshlaghy
2
(
Professor, Department of Industrial Management, Science and Research Unit, Islamic Azad University, Tehran, Iran
)
Morteza Mosakhani
3
(
Professor, Department of Management , Science and Research Branch, Islamic Azad University, Tehran, Iran
)
Keywords: Machine Learning, Users, Content, Social networks, EMOTIONAL ANALYSIS,
Abstract :
Recently,social networks have attracted special attention. In various social networks, users are constantly expressing their public as well as private opinions on various topics. Twitter is one of these social networks that has become very popular in the last decade. This social network provides organizations with a fast and effective way to analyze customers' feelings, views, and criticisms of market success. Emotional analysis is a process in which people's opinions, feelings, and attitudes about a particular subject are extracted. There has been a lot of research on emotion analysis based on user comments, documents and articles. Analysis of what is being said is very different from Twitter data, because Twitter tweets are limited to 280 characters and force users to express their feelings concisely. The best results in emotion classification are obtained from machine learning techniques such as simple Bayes and support vector machine. In this research, a method for analyzing emotions in social networks is presented. In this regard, we have tried to improve the classification of text by Bayesian method to some extent by focusing on the stages of data preprocessing and feature selection.users' feelings are analyzed. The classification problem has been formulated and solved using the latest achievements in the field of machine learning. . To evaluate the proposed method in this dissertation is from the Twitter data set scenario. The proposed method is compared with other classification methods. Has shown the best performance.
_||_
