Pratama, Muhammad Lutfi (2023) Analisis Performansi Naïve Bayes Classifier Dan Random Forest Terhadap Sentimen Kebijakan Kenaikan Harga Bbm Di Indonesia. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (Cover)
19081010049.-cover.pdf Download (583kB) | Preview |
|
|
Text (BAB 1)
19081010049.-bab1.pdf Download (126kB) | Preview |
|
![[img]](http://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (BAB 2)
19081010049.-bab2.pdf Restricted to Registered users only until 16 May 2025. Download (638kB) |
|
|
Text (BAB 3)
19081010049.-bab3.pdf Restricted to Registered users only until 16 May 2025. Download (604kB) |
|
|
Text (BAB 4)
19081010049.-bab4.pdf Restricted to Registered users only until 16 May 2025. Download (1MB) |
|
|
Text (BAB 5)
19081010049.-bab5.pdf Download (50kB) | Preview |
|
|
Text (Daftar pustaka)
19081010049.-daftarpustaka.pdf Download (183kB) | Preview |
|
|
Text (Lampiran)
19081010049.-lampiran.pdf Restricted to Registered users only until 16 May 2025. Download (136kB) |
Abstract
Fuel is a crucial commodity in people's economic activities. The fuel price increase policy can negatively affect the economic growth in community. However, the government make various good decisions, such as the BLT BBM. This phenomenon raises various sentiments in society. Knowing public sentiment can be a benchmark for the government in making decisions. Therefore, Naïve Bayes Classifier (NBC) and Random Forest (RF) algorithms to classify public sentiment towards the fuel price increase policy through Twitter text data, with 250 thousand tweet datasets. Sentiment class labels include positive, neutral, and negative. Performance analysis for each algorithm consider by accuracy, recall, and average the AUC-ROC score. Both algorithms will go through hyperparameter tuning process, for NBC that is the laplace smoothing value and for RF that is the minimum samples split and minimum samples leaf values. It was concluded that RF performance more useful as it reached 85.15% accuracy and 94.62% average AUC-ROC score. However, NBC reached accuracy value of 79.74% and the average AUC-ROC is 89.83%.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76.6 Computer Programming |
||||||||||||
| Divisions: | Faculty of Computer Science | ||||||||||||
| Depositing User: | Lutfi Muhammad Pratama | ||||||||||||
| Date Deposited: | 16 May 2023 08:17 | ||||||||||||
| Last Modified: | 16 May 2023 08:17 | ||||||||||||
| URI: | http://repository.upnjatim.ac.id/id/eprint/13258 |
Actions (login required)
 |
View Item |