Chest X-ray classification with CNNs
Developed and implemented a Convolutional Neural Network (CNN) model using PyTorch to classify chest X-ray images of 'normal' and 'pneumonia'. Leveraged techniques commonly used in object detection for efficient data preprocessing and dynamic feature engineering and achieved a validation accuracy of 86%.
View on GitHub
Melanoma Prediction
Developed a convolutional neural network (CNN) model for melanoma detection, achieving 90% accuracy. The model effectively differentiates between malignant and benign skin lesions, contributing to early diagnosis and treatment planning.
View on GitHub
Customer Churn Prediction(Telco)
Developed and implemented a deep neural network using PyTorch to predict customer churn from the Telco Customer Churn dataset. Conducted extensive exploratory data analysis and preprocessing, including handling missing values and converting categorical variables to numerical ones. Utilized techniques such as batch normalization and dropout for improved model performance. Achieved a training loss that steadily decreased over 100 epochs, indicating effective learning.
View on GitHub