metadata

language:
  - hu
  - sk
  - pl
  - cs
tags:
  - emotion-classification
  - roberta
  - fine-tuned
  - multilingual
license: mit
datasets:
  - custom
model-index:
  - name: Multilingual Fine-tuned RoBERTa for Emotion Classification
    results:
      - task:
          type: text-classification
          name: Multilingual Emotion Classification
        dataset:
          name: Multilingual Custom Dataset (Hungarian, Slovak, Polish, Czech)
          type: text
        metrics:
          - name: Precision (Macro Avg)
            type: precision
            value: 0.86
          - name: Recall (Macro Avg)
            type: recall
            value: 0.86
          - name: F1 Score (Macro Avg)
            type: f1
            value: 0.86
          - name: Accuracy
            type: accuracy
            value: 0.84

Multilingual Fine-tuned RoBERTa Model for Emotion Classification

Model Description

This model is a multilingual fine-tuned version of the RoBERTa model, specifically tailored for emotion classification tasks in Hungarian, Slovak, Polish, and Czech languages. The model was trained to classify textual data into six emotional categories (anger, fear, disgust, sadness, joy, and none of them).

Intended Use

This model is intended for classifying textual data into emotional categories across multiple languages, including Hungarian, Slovak, Polish, and Czech. It can be used in applications such as sentiment analysis, social media monitoring, customer feedback analysis, and similar tasks. The model predicts the dominant emotion in a given text among the six predefined categories.

Metrics

Class	Precision (P)	Recall (R)	F1-Score (F1)
anger	0.74	0.81	0.77
fear	0.98	0.98	0.98
disgust	0.94	0.95	0.95
sadness	0.87	0.87	0.87
joy	0.89	0.89	0.89
none of them	0.77	0.69	0.73
Accuracy			0.84
Macro Avg	0.86	0.86	0.86
Weighted Avg	0.84	0.84	0.84

Overall Performance

Accuracy: 0.84
Macro Average Precision: 0.86
Macro Average Recall: 0.86
Macro Average F1-Score: 0.86

Class-wise Performance

The model demonstrates strong performance across different emotional categories, with particularly high precision, recall, and F1 scores in the fear, disgust, and joy categories. The model performs moderately well in detecting anger and none of them categories, but still achieves adequate accuracy in these cases.

Limitations

Context Sensitivity: The model may struggle with recognizing emotions that require deeper contextual understanding.
Class Imbalance: The model's performance on the "none of them" category suggests that further training with more balanced datasets could improve accuracy.
Generalization: The model's performance may vary depending on the text's domain, language style, and length, especially across different languages.

How to Use

You can use this model directly with the transformers library from Hugging Face. Below is an example of how to load and use the model:

from transformers import pipeline

# Load the fine-tuned model
classifier = pipeline("text-classification", model="visegradmedia-emotion/Emotion_RoBERTa_pooled_V4")

# Example usage
result = classifier("Nagyon örömtelinek érzem magam ma!")
print(result)