Deep Learning Approach For Identification Of Poverty Through Sentiment Analysis

Redjeki, Sri and Widyarto, Setyawan and Suwasto, Anang (2020) Deep Learning Approach For Identification Of Poverty Through Sentiment Analysis. In: International Multidisciplinary Postgraduate Virtual Conference 2020 (IMPC20).

Text (Deep Learning Approach For Identification Of Poverty Through Sentiment Analysis) paper_IMPC20_UNISEL.pdf - Published Version Download (252kB)

Abstract

This research aims to identify poverty in Indonesia through sentiment analysis using a deep learning approach to the Long Short-Term Memory (LSTM) method. Poverty is one of the main problems that the Indonesian government has become aware of over the years. Many policies have been implemented and created by the government, either with their efforts or assistance from other countries or the World Bank. The dataset used is 10288 twitter data that is crawled using poverty-related keywords. Training uses 80% and testing uses 20 datasets. The training data is further divided into 2, namely the training set and the validation set. The LSTM model produces a training accuracy of 88% with a validation-accuracy of 72%. Keywords : Deep Learning; poverty; LSTM; sentiment analysis

Item Type:	Conference or Workshop Item (Paper)
Additional Information:	Penulis: Sri Redjeki, Setyawan Widyarto, dan Anang Suwasto
Uncontrolled Keywords:	Deep Learning; poverty; LSTM; sentiment analysis
Subjects:	A Karya Umum (General) > Ilmu Komputer (Computer Science) > Analisis Sistem A Karya Umum (General) > Ilmu Komputer (Computer Science) > Program Aplikasi
Divisions:	e-Artikel (e-Articles)
Depositing User:	Titis Pratiwi
Date Deposited:	27 Mei 2022 02:50
Last Modified:	27 Mei 2022 02:50
URI:	http://eprints.utdi.ac.id/id/eprint/9662

Actions (login required)

View Item