Brain Tumor Radiogenomic Classification Project This two notebooks were written to compete in the Kaggle machine learning competition.
In one notebook Neural Network various architectures are constructed, using model checkpoint and early stopping callbacks.
In the other notebook a different approach is used. Keras CNN pre-trained models are used for transfer learning, extracting features from the dataset. After the feature extracrion process, ensable voting is used among classical ML tequinques. Here a differenct image preprocessing and image augmentation are developed,