Vivalyse is an AI-powered ML model that assesses confidence and clarity in viva speeches using Natural Language Processing (NLP) and audio processing. From MFCC and text embeddings like BERT, GloVe, etc., it focuses on confidence and clarity for classification. The model ensures objective and fair evaluations applicable in education, HR, and AI-driven hiring.
- Programming Language: Python
- NLP Techniques: TF-IDF, BART, T5, GloVe
- Machine Learning Models: Random Forest, MLP, K-Means, Adaboost
- Audio Processing: MFCC, LPC
- Ensemble Learning: Stacking Models (Random Forest, MLP, KNN)
✅ Speech Feature Extraction: Extracts MFCC and LPC features to analyze tone, pace, and clarity.
✅ ML-based Classification: Uses stacking models for accurate confidence and clarity classification.
✅ High Performance: Achieves 92.8% F1-score for confidence classification and 91.4% for clarity.
✅ Bias Reduction System: Ensures fair assessments with applications in education, HR, and AI-driven hiring.
- Clone the repository:
git clone https://github.com/SaiSrujanReddyP/MachineLearning cd MachineLearning - Install dependencies:
pip install -r requirements.txt
- Run the main script:
python main.py
This project is licensed under the MIT License - see the LICENSE file for details.