This repository focuses on sentiment analysis for user reviews of various banking applications. By leveraging Natural Language Processing (NLP) techniques and Machine Learning models, the goal is to classify and understand user sentiments, providing actionable insights to improve application performance.
Banking applications often receive feedback from users that reflects their experience and satisfaction. This project analyzes these user reviews to classify sentiments into:
- Positive
- Neutral
- Negative
The insights from this analysis can help improve:
- User satisfaction
- Application functionality
- Banking services optimization
The dataset consists of user reviews from various banking applications. Below is the structure of the dataset:
| Column | Description |
|---|---|
at |
Timestamp of the review |
userName |
The user's name |
content |
The review text |
score |
Review rating (1-5) |
application |
Name of the banking application |
- Number of Reviews: 6,170
- Applications Covered: Multiple banking apps
📂 The dataset is stored in the file: dataPerbankan.csv
The project utilizes multiple machine learning models to classify user sentiments. These include:
- 🔍 K-Nearest Neighbors (KNN)
- 🌲 Random Forest
- 🧠 Naive Bayes
- 🔗 Support Vector Machine (SVM)
The models are evaluated using:
- Accuracy
- Precision
- Recall
- F1-Score
Key insights are visualized through:
- Confusion Matrices
- Performance Charts
- Data Distribution Plots