Neural-Network-Implementations

Implementing following Neural Networks

• Multi Layer Perceptron
• Radial Basis Function Neural Network
• Stacked Autoencoders
• Extreme Learning Machines
• Deep Neural Network (Autoencoder + ELM)
• 1D Convolutional Neural Network

Data (except CNN)

• Data has 72 features
• Output is binary and unidimentional
• Hold-out Cross validation is used for evaluation

Data (for CNN)

• This has the ECG data with 1000 features.
• Target field is unidimensional and binary.
• Total there are 1000 records.
• Keras is used for the implementation.