/Prediction_with_SVM_on_Tumor_Dataset

Here classification on the tumor dataset is performed using Support Vector Machine. Tumor dataset has nine features and a binary output. Nine features correspond to measurements obtained from medical imaging data. The output labels are logic 1 or 0 and corresponds to the presence or absence of tumor

Primary LanguagePython

Prediction with SVM on Tumor Dataset

Here classification on the tumor dataset is performed using Support Vector Machine. Tumor dataset has nine features and a binary output. Nine features correspond to measurements obtained from medical imaging data. The output labels are logic 1 or 0 and corresponds to the presence or absence of tumor.

Dataset

Dataset consists of training dataset, labels for training dataset and test dataset as the text documents.

Requirements

  • Python 3.5 or higher
  • Numpy
  • Sklearn

Results

  • Extracted tumor dataset with input having nine distinct features, binary labeled output and applied SVM classifier.
  • Used Linear, Polynomial & RBF Kernels to improve accuracy and chose the best model by tuning Hyperparameters.
  • Successfully ranked in top 28% in Kaggle competition (out of 170 participants) to predict presence of tumor.