Personal project

Indonesian Sentiment Analysis

A hybrid IndoBERTweet-BiLSTM model for sentiment classification in Indonesian tweets, reviews, and public opinions.

Natural Language ProcessingDeep learning project

Explore Sentiment classification evaluation

IndoBERTweet and BiLSTM classification chart: accuracy 0.9004, precision 0.9014, recall 0.9004, and F1-score 0.9003

Tech Stack

PythonPyTorchHugging Face TransformersIndoBERTweetBiLSTMGoogle Colab

Feature spotlights / Indonesian Sentiment Analysis

Sentiment classification evaluation

The experiment combines IndoBERTweet and BiLSTM for Indonesian sentiment analysis. These two views document its reported metrics and place them alongside the other models included in the comparison.

Choose a section to explore

Sentiment classification evaluation

IndoBERTweet + BiLSTM results

The recorded classification results show accuracy of 0.9004, precision of 0.9014, recall of 0.9004, and an F1-score of 0.9003. The chart brings the four metrics together for the hybrid IndoBERTweet-BiLSTM model.

90.04% accuracy in the reported experiment, with precision, recall, and F1 shown alongside it.

Challenge

Informal Indonesian text requires models that capture contextual meaning and sequential dependencies.

Capture sentiment in informal text through contextual and sequence-level information.

Approach

Combined IndoBERTweet embeddings with BiLSTM and evaluated predictions using classification metrics and a confusion matrix.

Explores Indonesian sentiment classification with contextual embeddings and bidirectional sequence modeling.

Implementation

IndoBERTweet contextual embeddings
Bidirectional LSTM sequence modeling
Sentiment classification
Accuracy, precision, recall, and F1 evaluation

Features

Indonesian sentiment classification
Combined tweet, review, and public opinion datasets
Contextual language embeddings
Bidirectional sequence modeling
Classification metrics and confusion matrix

Highlights

Hybrid Indonesian sentiment model
Experiments on combined Indonesian text datasets
PyTorch and Hugging Face development

Let's talk about your next project.

Let's Talk

Loading page...

Back to projects

Personal project

Indonesian Sentiment Analysis

A hybrid IndoBERTweet-BiLSTM model for sentiment classification in Indonesian tweets, reviews, and public opinions.

Natural Language ProcessingDeep learning project

Explore Sentiment classification evaluation

Tech Stack

PythonPyTorchHugging Face TransformersIndoBERTweetBiLSTMGoogle Colab

Feature spotlights / Indonesian Sentiment Analysis

Sentiment classification evaluation

The experiment combines IndoBERTweet and BiLSTM for Indonesian sentiment analysis. These two views document its reported metrics and place them alongside the other models included in the comparison.

Choose a section to explore

Sentiment classification evaluation

IndoBERTweet + BiLSTM results

90.04% accuracy in the reported experiment, with precision, recall, and F1 shown alongside it.

Challenge

Informal Indonesian text requires models that capture contextual meaning and sequential dependencies.

Capture sentiment in informal text through contextual and sequence-level information.

Approach

Combined IndoBERTweet embeddings with BiLSTM and evaluated predictions using classification metrics and a confusion matrix.

Explores Indonesian sentiment classification with contextual embeddings and bidirectional sequence modeling.

Implementation

IndoBERTweet contextual embeddings
Bidirectional LSTM sequence modeling
Sentiment classification
Accuracy, precision, recall, and F1 evaluation

Features

Indonesian sentiment classification
Combined tweet, review, and public opinion datasets
Contextual language embeddings
Bidirectional sequence modeling
Classification metrics and confusion matrix

Highlights

Hybrid Indonesian sentiment model
Experiments on combined Indonesian text datasets
PyTorch and Hugging Face development

Let's talk about your next project.

Let's Talk