Nusabert Base Indonesian Plutchik Emotion Analysis V2

🚀 EmoSense-ID: Indonesian Emotion Classifier

EmoSense-ID is a powerful model designed to precisely identify and analyze emotions in Indonesian texts. It leverages Plutchik's eight fundamental emotions, offering valuable insights into users' emotional responses on social media.

🚀 Quick Start

The widget section below provides real - world examples of how the EmoSense - ID model analyzes different Indonesian texts and outputs the corresponding emotional scores.

Example 1

• Text: How irresponsible it is to just throw garbage into the river like that. There are environmental activists who even end up in jail for speaking out about environmental concerns. This unemployed person actually finds it okay to litter randomly. No wonder you're having a hard time; your behavior is really bothering others.
• Output:

  Label	Score
  Disgust	0.672
  Anger	0.282
  Sadness	0.033
  Joy	0.004
  Surprise	0.003
  Trust	0.003
  Fear	0.002
  Anticipation	0.001

Example 2

• Text: In February 2009, a journalist from Jawa Pos Radar Bali was brutally murdered because of a corruption report. In January 2019, the President granted clemency to the mastermind of the Prabangsa murder, reducing the sentence from life imprisonment to just 20 years. What a painful step backward!
• Output:

  Label	Score
  Sadness	0.604
  Anger	0.194
  Surprise	0.127
  Joy	0.021
  Fear	0.018
  Disgust	0.018
  Anticipation	0.016
  Trust	0.003

Example 3

• Text: I'm really in awe of their life journey. If it's made into a movie, it'll definitely be a hit and super inspiring.
• Output:

  Label	Score
  Joy	0.9637
  Trust	0.0219
  Anticipation	0.0079
  Surprise	0.0029
  Disgust	0.0013
  Sadness	0.0010
  Anger	0.0007
  Fear	0.0006

Example 4

• Text: I swear they must be freed!!! Why don't you just arrest the corruptors instead of dealing with AI memes? I'm so angry!
• Output:

  Label	Score
  Anger	0.9889
  Disgust	0.0035
  Sadness	0.0026
  Fear	0.0015
  Surprise	0.0012
  Trust	0.0011
  Anticipation	0.0009
  Joy	0.0003

Example 5

• Text: I've never been in a relationship. Now I'm so lonely. I want my friends to introduce me to a guy, but they're also struggling. I really regret not dating during school and college.
• Output:

  Label	Score
  Sadness	0.9526
  Anger	0.0175
  Fear	0.0114
  Disgust	0.0079
  Trust	0.0038
  Anticipation	0.0036
  Joy	0.0019
  Surprise	0.0013

Example 6

• Text: The Indonesian Broadcasting Commission (KPI) has requested that television broadcasts present a positive and educational image of the Indonesian National Police. This was stated by the Chairman of the Central KPI, Ubaidillah, in a panel discussion.
• Output:

  Label	Score
  Anticipation	0.4323
  Trust	0.3996
  Joy	0.0500
  Anger	0.0388
  Disgust	0.0362
  Surprise	0.0186
  Fear	0.0137
  Sadness	0.0108

📚 Documentation

Model Details

Model Description

The EmoSense - ID is a model crafted to identify and analyze emotions in Indonesian texts based on Plutchik's eight basic emotions: Anticipation, Anger, Disgust, Fear, Joy, Sadness, Surprise, and Trust. This model is developed using the [NusaBERT - base](https://huggingface.co/LazarusNLP/NusaBERT - base) and trained using Indonesian tweets categorized into eight emotion categories. The evaluation results of this model can be utilized to analyze emotions in social media, providing insights into users' emotional responses.

Bias

Keep in mind that this model is trained using certain data which may cause bias in the emotion classification process. Therefore, it is important to consider and account for such biases when using this model.

Evaluation Results

The model was trained using the Hyperparameter Tuning technique with Optuna. In this process, Optuna conducted five trials to determine the optimal combination of learning rate (1e - 6 to 1e - 4) and weight decay (1e - 6 to 1e - 2). Each trial trained the BERT model with different hyperparameter configurations on the training dataset and then evaluated using the validation dataset. After all the experiments are completed, the best hyperparameter combination is used to train the final model.

📄 License

This project is licensed under the MIT license.

📚 Citation

@misc{Ardiyanto_Mikhael_2024,
    author    = {Mikhael Ardiyanto},
    title     = {EmoSense-ID},
    year      = {2024},
    URL       = {Aardiiiiy/EmoSense-ID-Indonesian-Emotion-Classifier},
    publisher = {Hugging Face}
}