A Method to Identify the Current Mood of Social Media Users

Mood of a person changes frequently during a day. A mood can be categorized as happy, sad, calm and angry. Most of the people today share their daily activities, opinions, feelings regularly in social media. Identification of current mood will be useful for recommendation systems to change or elevat...

Full description

Saved in:
Bibliographic Details
Published in2019 14th Conference on Industrial and Information Systems (ICIIS) pp. 356 - 359
Main Authors Nimeshika, Supuni, Ahangama, Supunmali
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2019
Subjects
Convolution Neural Network
Current Mood
Emoticons
Natural Language Processing
OCR
Online AccessGet full text

Cover

More Information
Summary:Mood of a person changes frequently during a day. A mood can be categorized as happy, sad, calm and angry. Most of the people today share their daily activities, opinions, feelings regularly in social media. Identification of current mood will be useful for recommendation systems to change or elevate moods. Thus, this proposed system identifies the current mood of a person by mining their social media contents such as posts, comments, image posts and emoticons. In the proposed solution, a score for the current mood is calculated in two stages; a score for text contents (images with text and posts/comments) and score for emoticons were computed. Posts made within a 24-hour period will be considered for the current mood and scores for multiple posts are combined using a temporal weighted average. Text classification is done using a 1D Convolution Neural Network and emoticon classification is performed using a survey. Finally, an overall accuracy of 85% is achieved.
DOI:10.1109/ICIIS47346.2019.9063291