Early detection of anxiety in social media using Convolution neural network / Mohd Tharwan Hadi Ruslan

Ruslan, Mohd Tharwan Hadi (2021) Early detection of anxiety in social media using Convolution neural network / Mohd Tharwan Hadi Ruslan. Degree thesis, Universiti Teknologi MARA, Terengganu.

Abstract

With pandemic covid-19 getting worse and most people need to be stayed at home, will surely make rate of mental health conditioner to be increased. Mental health shouldn't be taken lightly since the sole reason of suicide come from underlying mental health problems. But since most people need to stay at home, then by no mean detecting mental health patient will be difficult. So, the only way left to detect anxiety patients is through online or specifically social media. Since anxiety is the leading condition among all mental health condition, beating depression even, then this project will focus on detecting anxiety patient through Twitter tweets. The reason why this project will use social media Twitter to detect anxiety is because these anxiety conditioners sometime will post that they having anxiety or nearly having anxiety on social media. So this project aim to detect anxiety early through social media Twitter. This project surely be able to help the society, by giving mental health facility the capability to detect anxiety patient through Twitter, contact them early, and giving them early treatments. The system will get the data from the user, which in this case is mental health staff, and then decide whether the data received is positive anxiety or not. This project will consist of literature study, data collection, data analysis, data cleaning, user interface design and classifier design. From literature study, the CNN algorithm had been chosen for this project, since CNN algorithm proof to be the best algorithm for classifying anxiety patients. Twitter API and tweepy will be used for data collection and textblob for sentiment analysis. Then there will be a lot of preprocess step to clean the data. As for the classifier design, the keras function will be used to generate CNN classifier. And python GUI for user interface. Model and system evaluation are also done for the project, proof that the classifier and user interface able to function as intended. In conclusion, this project may help decreasing the mental health patients, hence decrease the suicide rate.

Metadata

Item Type: Thesis (Degree)
Creators:
Creators
Email / ID Num.
Ruslan, Mohd Tharwan Hadi
UNSPECIFIED
Contributors:
Contribution
Name
Email / ID Num.
Thesis advisor
Jantan, Hamidah
UNSPECIFIED
Subjects: Q Science > QA Mathematics > Instruments and machines
Q Science > QA Mathematics > Instruments and machines > Electronic Computers. Computer Science
Q Science > QA Mathematics > Instruments and machines > Electronic Computers. Computer Science > Neural networks (Computer science)
Q Science > QA Mathematics > Analytic mechanics
Divisions: Universiti Teknologi MARA, Terengganu > Dungun Campus > Faculty of Computer and Mathematical Sciences
Programme: Bachelor of Computer Science (Hons)
Keywords: Pandemic Covid-19 ; Mental Health ; Early Detection Of Anxiety ; Convolution Neural Network
Date: February 2021
URI: https://ir.uitm.edu.my/id/eprint/55140
Edit Item
Edit Item

Download

[thumbnail of 55140.pdf] Text
55140.pdf

Download (121kB)

Fulltext

Fulltext is available at:
  • Kaunter Perkhidmatan Maklumat | Perpustakaan Cendekiawan | Dungun

ID Number

55140

Indexing

|

Statistic

Statistic details