Analysis of Users’ Opinions about Reasons for Divorce

One of the most important issues related to knowledge discovery is the field of comment mining. Opinion mining is a tool through which the opinions of people who comment about a specific issue can be evaluated in order to achieve some interesting results. This is a subset of data mining. Opinion min...

Full description

Bibliographic Details
Main Authors: Fatemeh Eghrari Solout, Mehdi Hosseinzadeh
Format: Article
Language:English
Published: Science and Research Branch,Islamic Azad University 2016-12-01
Series:Journal of Advances in Computer Engineering and Technology
Subjects:
Online Access:http://jacet.srbiau.ac.ir/article_9552_fce4a3e7036a24d1b352810dd83e1796.pdf
Description
Summary:One of the most important issues related to knowledge discovery is the field of comment mining. Opinion mining is a tool through which the opinions of people who comment about a specific issue can be evaluated in order to achieve some interesting results. This is a subset of data mining. Opinion mining can be improved using the data mining algorithms. One of the important parts of opinion mining is the sentiment analysis in social networks. Today, the social networks contain billions of users' comments about different issues. In previous researches in this area, various methods have been used for Persian comments analysis. In these studies, preprocessing is one of the most important parts. It arranges the data set for analysis in a standard form. The number of hashtags selected for analysis is limited. To detect the positive and negative comments, knowledge extraction or neural network techniques have been used. The current research presents a method of analysis which can analyze any hashtag for each group of users and has no limitations in this regard. Type of hashtag, the number of likes, type of user and type of positive and negative sentences can be analyzed by this method. The results of simulation and comparison of divorce data set show that the proposed method has an acceptable performance.
ISSN:2423-4192
2423-4206