DL ICP1 - koushikskr/python GitHub Wiki

Introduction: This ICP is about predicting the diabetes disease in Deep Learning programming on Keras.

Question 1: Add more Dense layers to the existing code and check how the accuracy changes

Procedure: For the given code, added one more dense layer with input argument 60. Could see the increased in mapping parameters value 1260. Loss score became 0.55 and accuracy is 0.72

Question 2: Change the data source to Breast Cancer data set * available in the source folder and make required changes

Procedure: Read the given csv file. split the data set with target column 'diagnosys' and created data. split the data into train and test data. Used label encoder to change the char target column value to numeric. Now created the Sequential() instance and added hidden layers to it. Fit the model with trained data. Found out the summary and evaluation Loss score is 0.31 and accuracy is 0.91

Question 3: Normalize the data before feeding the data to the model and check how the normalization change your accuracy

Procedure: Followed the same procedure as question 2. Before sending the trained and test data to model created standard scalar object and scaled the data using it. Observed the loss score is 0.13 and accuracy is 0.96