Please use this identifier to cite or link to this item:
https://repositori.uma.ac.id/handle/123456789/24547
Title: | Analisis Sentimen pada Media Sosial dengan Menggunakan Algoritma Radial Basis Function Neural Networks (RBFNN) |
Other Titles: | Sentiment Analysis on Social Media Using the Radial Basis Function Neural Networks (RBFNN) Algorithm |
Authors: | Rian, Ade |
metadata.dc.contributor.advisor: | Syah, Rahmad |
Keywords: | analisis sentimen;algoritma radial basis function neural network;twitter;k-fold cross validation;sentiment analysis |
Issue Date: | 13-Jun-2024 |
Publisher: | Universitas Medan Area |
Series/Report no.: | NPM;188160004 |
Abstract: | Analisis sentimen atau opinion mining merupakan prosedur mengetahui, mengekstraksi, dan mengolah informasi teks secara otomatis untuk memperoleh data sentimen yang terdapat pada kalimat opini. Prosedur analisis data bisa dilakukan melewati teknik machine learning. Pada penelitian ini, analisis sentimen dilakukan guna mengetahui pemahaman ataupun trend persepsi kepada perbankan syariah. Analisis sentimen membantu mengidentifikasi komentar atau pendapat yang mempunyai kecenderungan sentimen negatif atau positif, yang dapat dijadikan acuan untuk pembaruan layanan. Dalam penelitian ini, data yang diperoleh sebanyak 1837 dengan membagi 1768 positif dan 69 negatif dari Twitter dengan teks berbahasa Indonesia yang terkait dengan bank syariah Indonesia. Hasilnya dengan menggunakan proporsi data training 60% dan testing 40% menghasilkan performa akurasi dengan algoritma RBFNN sebesar 96,8%, presisi 96,8%, dan recall 96,8%, selanjutnya performa akurasi dengan K-Fold Cross Validation sebesar 96,2%. Proporsi data training 70% dan testing 30% menghasilkan performa akurasi dengan algoritma RBFNN sebesar 96,3%, presisi 96,3%, dan recall 96,3%, selanjutnya performa akurasi dengan K-Fold Cross Validation sebesar 96,2%. Kemudian menggunakan proporsi data training 80% dan testing 20% menghasilkan performa akurasi dengan algoritma RBFNN sebesar 97%, presisi 97%, dan recall 97%, selanjutnya performa akurasi dengan K-Fold Cross Validation sebesar 96,2%. Sentiment analysis or opinion mining is a procedure for knowing, extracting and processing text information automatically to obtain sentiment data contained in opinion sentences. Data analysis procedures can be carried out using machine learning techniques. In this research, sentiment analysis was carried out to determine understanding or trends in perceptions of Islamic banking. Sentiment analysis helps identify comments or opinions that tend to have negative or positive sentiment, which can be used as a reference for service updates. In this research, the data obtained was 1837 by dividing 1768 positive and 69 negative from Twitter with Indonesian text related to Indonesian sharia banks. The results using a training data proportion of 60% and testing 40% produced accuracy performance with the RBFNN algorithm of 96.8%, precision of 96.8%, and recall of 96.8%, then accuracy performance with K-Fold Cross Validation was 96.2 %. The proportion of training data of 70% and testing of 30% produces accuracy performance with the RBFNN algorithm of 96.3%, precision of 96.3%, and recall of 96.3%, then accuracy performance with K-Fold Cross Validation of 96.2%. Then using a training data proportion of 80% and testing 20% produces accuracy performance with the RBFNN algorithm of 97%, precision of 97%, and recall of 97%, then accuracy performance with K-Fold Cross Validation of 96.2%. |
Description: | 53 Halaman |
URI: | https://repositori.uma.ac.id/handle/123456789/24547 |
Appears in Collections: | SP - Informatic Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
188160004 - Ade Rian - Fulltext.pdf | Cover, Abstract, Chapter I, II, III, V, Bibliography | 1.73 MB | Adobe PDF | View/Open |
188160004 - Ade Rian - Chapter IV.pdf Restricted Access | Chapter IV | 428.35 kB | Adobe PDF | View/Open Request a copy |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.