PANJAITAN, LASTRI RAHELITA and Utami, Alvi Syahrini and Rachmatullah, M. Naufal (2024) ANALISIS SENTIMEN TWEET TERHADAP PENGGUNAAN CHATGPT SEBAGAI ASISTEN VIRTUAL MENGGUNAKAN LONG SHORT-TERM MEMORY. Undergraduate thesis, Sriwijaya University.
Text
RAMA_55201_09021282025083.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (6MB) | Request a copy |
|
Text
RAMA_55201_09021282025083_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (8MB) | Request a copy |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (823kB) |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (337kB) | Request a copy |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (163kB) | Request a copy |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (671kB) | Request a copy |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (9kB) | Request a copy |
|
Text
RAMA_55201_09021282025083_0022127804_0001129204_07_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (154kB) | Request a copy |
Abstract
Sentiment analysis is a research field that is increasingly gaining popularity with the growing number of internet users and the availability of online text data. However, the abundance of text data also poses challenges in conducting sentiment analysis. Datasets containing long and complex text documents require suitable methods. In this study, researchers utilized Long Short-Term Memory (LSTM) as the sentiment analysis method. The dataset used consisted of 106.695 ChatGPT tweets, obtained from Kaggle, and was divided into 80% training data and 20% test data. GloVe was employed as the word embedding technique in this research. The aim of this study was to measure the performance of the LSTM modelPEN in classifying ChatGPT user sentiments. Researchers conducted manual hyperparameter tuning with 10 experiments for each hyperparameter, selecting the best results for the LSTM model configuration. The findings revealed that the LSTM model with a dropout rate of 0.3, 64 LSTM units, recurrent dropout on the LSTM layer at 0.3, 20 epochs, and a batch size of 128 achieved the highest accuracy, reaching 88, 86%. This configuration resulted in an precision, recall, and F1-score of 90% in sentiment analysis.
Item Type: | Thesis (Undergraduate) |
---|---|
Uncontrolled Keywords: | ASISTEN VIRTUAL, LONG SHORT-TERM MEMORY |
Subjects: | T Technology > T Technology (General) > T1-995 Technology (General) T Technology > TA Engineering (General). Civil engineering (General) > TA165 Engineering instruments, meters, etc. Industrial instrumentation T Technology > TA Engineering (General). Civil engineering (General) > TA213-215 Engineering machinery, tools, and implements |
Divisions: | 09-Faculty of Computer Science > 55201-Informatics (S1) |
Depositing User: | Lastri Rahelita Panjaitan |
Date Deposited: | 26 Feb 2024 01:49 |
Last Modified: | 26 Feb 2024 01:49 |
URI: | http://repository.unsri.ac.id/id/eprint/141047 |
Actions (login required)
View Item |