Skip to content

molcik/python-neuron

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Neuron class

Neuron class provides LNU (Linear Neural Unit), QNU (Quadratic Neural Unit), RBF (Radial Basis Function), MLP (Multi Layer Perceptron), MLP-ELM (Multi Layer Perceptron - Extreme Learning Machine) neurons learned with Gradient descent or LeLevenberg–Marquardt algorithm. This class is suitable for prediction on time series.

Dependencies

Neuron class needs pandas and numpy to work propertly.

Example of usage

Consider Y are targets and X are inputs.

## LNUGD

neuron = LNUGD()
prediction = 1
yn, w, e, Wall, MSE = neuron.train(Y_train, X_train, epochs=2, prediction=prediction)
yn, w, Wall, MSE, e = neuron.countSerie(Y, X, logging=False, prediction=prediction)

QNULM

neuron = QNULM()
prediction = 1
yn, w, e, Wall, MSE = neuron.train(Y_train, X_train, epochs=10, prediction=prediction)
yn, w, MSE, e = neuron.countSerie(Y, X, logging=False, prediction=prediction)

RBF

neuron = RBF()
prediction = 1
neuron.train(Y_train, X_train, prediction=prediction)
yn = neuron.count(Y,X, logging=True, beta=0.01, prediction=prediction)

MLPGD

neuron = MLPGD()
prediction = 1
yn = neuron.count(Y_train, X_train, prediction=prediction, epochs=5)
yn = neuron.count(Y, X, prediction=prediction, epochs=1)

MLPELM

neuron = MLPELM()
prediction = 1
yn = neuron.count(Y_train, X_train, prediction = prediction, epochs = 10)
yn = neuron.count(Y, X, prediction = prediction)

MLPLMWL

neuron = MLPLMWL()
prediction = 1
yn = neuron.count(Y, X, learningWindow = 50, overLearn = 10,  prediction = prediction)

Support me

If you find this useful, consider supporting independent open-source development and buy me a coffee.

buy me a coffee