%---------------------------------------- Data ------------------------------------------%
In neuralnet.ipynb we consider the data of people applying for credit cards, and a banker has placed them in one of the following categories:
Approved (1)
Denied (-1)
Each data point has 16 associated features (found in credit_data.csv) and a corresponding label (found in credit_label.csv). (# data = m = 653, # features = n = 16)
%------------------------------------ Goal of the project --------------------------------------%
The goal in this program is to create an artificial neural network (ANN) model with 3 layers.
Layer 1: 30 neurons and linear transfer function for the activation function.
Layer 2: 30 neurons and tanh function for the activation function.
Layer 3: (output layer) 2 neurons and softmax for the activation function.
We consider the data for two cases: data is unchanged, and data is normalized by the maximum value of each feature.
%------------------------------------ Required Packages --------------------------------------%
tensorflow (this notebook is compatible with v1, caution if you have v1 or v2 downloaded)
numpy
matplotlib
scipy
sklearn