Philliams / PySINN

Shepard Interpolation Neural Network - Theano and Keras implementation

PySINN

PySINN is a python implementation of the Shepard Interpolation Neural Network using Keras and Theano.

Dependencies:

PySINN uses Keras and Theano for all of the Machine Learning computations and Scikit Learn for the K-means clustering.

Usage:

# create Shepard Interpolation Neural Network object
s = SINN_model(data_training,labels_training,number_of_clusters,nodes_per_cluster,batch_size,epochs,verbose=1,model_optimizer='RMSprop',loss_metric='categorical_crossentropy')

# Train the model
s.train_model(data_training,labels_training,data_testing,labels_testing)

# Use model
predictions = s.predict(data_testing)