dnn_data_classifier

DNN Cancer Classifier

This section demonstrates a Deep Neural Network implementation to classify breast cancer tumours as benign or malignant depending on measurements taken directly from tumours.

Usage

1. Change directory to the dnn module:

   $ cd dnn_data_classifier
   

2. Run the application:

   $ python main.py
   

3. Any existing model will be removed and a new model will be trained. The parameters for training (epochs, steps, model directory, etc) can be altered in dnn_data_classifier/main.py. The expected output should be:

   Deleting resource: Model directory [nn_classifier].
   Removed resource: Model directory [nn_classifier].
   Deleting resource: Data [training_set.csv].
   Removed resource: Data [training_set.csv].
   Deleting resource: Data [test_set.csv].
   Removed resource: Data [test_set.csv].
   Model trained after 2000 steps.
   
   Model Accuracy: 0.927007
   

4. With the model now trained and fairly accurate, it can be used for prediction, either by entering your own measurements or by similating fake data:

   Predict classification: Enter own data (1) or simulate fake data (2)?
    Enter 1 or 2: 1
   Enter value 0-10 for clump_thickness: 1
   Enter value 0-10 for unif_cell_size: 5
   Enter value 0-10 for unif_cell_shape: 6
   Enter value 0-10 for marg_adhesion: 4
   Enter value 0-10 for single_epith_cell_size: 8
   Enter value 0-10 for bare_nuclei: 2
   Enter value 0-10 for bland_chrom: 6
   Enter value 0-10 for norm_nucleoli: 9
   Enter value 0-10 for mitoses: 3
   
   Class Prediction: malignant
   
   Would you like to try another prediction? Enter Y/N: y
   Predict classification: Enter own data (1) or simulate fake data (2)?
    Enter 1 or 2: 2
   Data generated:
   clump_thickness: 0
   unif_cell_size: 1
   unif_cell_shape: 2
   marg_adhesion: 3
   single_epith_cell_size: 4
   bare_nuclei: 5
   bland_chrom: 6
   norm_nucleoli: 7
   mitoses: 8
   
   Class Prediction: benign