AI / MLFeatured
Image Classification & Sentiment Analysis
Dicoding course projects Individual Coursework
Dicoding course projects covering a six-class image CNN and sentiment-analysis experiments with neural and linear models.
PythonTensorFlow / Kerasscikit-learnNumPypandas
What was broken?
These course projects explored how to prepare image and text data, train suitable models, and evaluate the results on held-out data.
How I fixed it
I prepared data and trained a CNN for six-class image classification. For sentiment analysis, I experimented with BiLSTM, Conv1D, and TF-IDF with LinearSVC, including preprocessing and held-out evaluation.
What it can do
- Six-class image classification with a CNN
- Sentiment analysis with BiLSTM and Conv1D
- TF-IDF and LinearSVC comparison
- Data preparation, training, and held-out evaluation
The stack that shipped it
TensorFlow/Keras supported neural-network experiments, while scikit-learn provided the TF-IDF and LinearSVC workflow. NumPy and pandas supported data preparation.
What changed
Completed as Dicoding course projects. No numerical accuracy, certification, public repository, or deployment is claimed here; the experiment artifacts are not available in this portfolio.