Emotion_predictor

Inspiration

I wanted to build a simple AI/ML project that works on real-world text data. Emotion classification felt interesting because it combines NLP with machine learning in a very practical way.

How I Built It

• Used a labelled text dataset containing different emotions.
• Cleaned and prepared the data.
• Converted text into numerical features using TF-IDF (Term Frequency-Inverse Document Frequency).
• Trained a Naive Bayes classifier to predict the correct emotion.
• Evaluated the model using accuracy and classification report.
• Added a function to test the model with custom sentences.

What I Learned

• Basics of NLP and text preprocessing.
• How TF-IDF works for feature extraction.
• Building and training a machine learning model in Python.
• How to evaluate ML models.
• Handling dataset issues, debugging, and running code in VS Code.

Challenges

• Incorrect column names in the dataset caused errors.
• Handling file paths and local CSV loading.
• Improving model accuracy with limited preprocessing.
• Understanding how train/test split impacts results.

Math Behind It (Optional LaTeX)

This is the basic formula used by the Naive Bayes classifier: $$P(\text{label} \mid \text{text}) = \frac{P(\text{text} \mid \text{label}) \cdot P(\text{label})}{P(\text{text})}$$

Built With

• ai
• data
• dataset
• machine-learning
• ml
• sklearn