Welcome to Portfolio

Food Vision Pro is a Streamlit app built with TensorFlow and CNN architecture, leveraging EfficientNet for deep learning-based food image classification. The model is fine-tuned on the Food101 dataset using mixed precision training and data augmentation techniques to accurately identify food items. It also integrates the NutritionixAPI for fetching

Credit Score Prediction is a machine learning project that predicts credit scores ('Good', 'Standard', 'Poor') using a streamlined pipeline. The project involves data extraction, cleaning, and preprocessing. Key techniques include Mutual Information for feature selection, PCA for dimensionality reduction, and XGBoost for accurate model training, ensuring reliable and robust predictions.

SMS Spam Classification is a machine learning project that classifies SMS messages as spam or not spam. It includes the entire pipeline from data collection and preprocessing to model training and evaluation, using Streamlit for an interactive user interface, ensuring efficient and accurate classification of messages.

Hate speech detection using a Decision Tree algorithm classifies text as hateful or non-hateful by evaluating features such as words and phrases. The algorithm builds a tree of decision rules, splitting nodes based on feature values, to predict the category of text, aiding in identifying and moderating harmful content.

MVP approach involves creating a simple model with basic features and minimal preprocessing. This allows for rapid testing, gathering feedback, and iterative improvements based on performance and user insights. Applied on Linear Regression.