Documentation
¶
Overview ¶
Neural provides struct to represents most common neural networks model and algorithms to train / test them.
Neural provides struct to represents most common neural networks model and algorithms to train / test them.
Neural provides struct to represents most common neural networks model and algorithms to train / test them.
Neural provides struct to represents most common neural networks model and algorithms to train / test them.
Neural provides struct to represents most common neural networks model and algorithms to train / test them.
Index ¶
- Constants
- func Accuracy(actual []float64, predicted []float64) (int, float64)
- func BackPropagate(mlp *MultiLayerNetwork, s *Pattern, o []float64, options ...int) (r float64)
- func ElmanTrain(mlp *MultiLayerNetwork, patterns []Pattern, epochs int)
- func Execute(mlp *MultiLayerNetwork, s *Pattern, options ...int) (r []float64)
- func HeavysideTransfer(d float64) float64
- func HeavysideTransferDerivate(d float64) float64
- func HyperbolicTransfer(d float64) float64
- func HyperbolicTransferDerivate(d float64) float64
- func MLPTrain(mlp *MultiLayerNetwork, patterns []Pattern, mapped []string, epochs int)
- func Predict(neuron *NeuronUnit, pattern *Pattern) float64
- func RandomNeuronInit(neuron *NeuronUnit, dim int)
- func RawExpectedConversion(patterns []Pattern) []string
- func SigmoidalTransfer(d float64) float64
- func SigmoidalTransferDerivate(d float64) float64
- func TrainNeuron(neuron *NeuronUnit, patterns []Pattern, epochs int, init int)
- func UpdateWeights(neuron *NeuronUnit, pattern *Pattern) (float64, float64)
- type MultiLayerNetwork
- type NeuralLayer
- type NeuronUnit
- type Pattern
Constants ¶
const (
SCALING_FACTOR = 0.0000000000001
)
Variables ¶
This section is empty.
Functions ¶
func Accuracy ¶
Accuracy calculate percentage of equal values between two float64 based slices. It returns int number and a float64 percentage value of corrected values.
func BackPropagate ¶ added in v0.2.0
func BackPropagate(mlp *MultiLayerNetwork, s *Pattern, o []float64, options ...int) (r float64)
BackPropagation algorithm for assisted learning. Convergence is not guaranteed and very slow. Use as a stop criterion the average between previous and current errors and a maximum number of iterations. [mlp:MultiLayerNetwork] input value [s:Pattern] input value (scaled between 0 and 1) [o:[]float64] expected output value (scaled between 0 and 1) return [r:float64] delta error between generated output and expected output
func ElmanTrain ¶ added in v0.2.0
func ElmanTrain(mlp *MultiLayerNetwork, patterns []Pattern, epochs int)
ElmanTrain train a mlp MultiLayerNetwork with BackPropagation algorithm for assisted learning.
func Execute ¶ added in v0.2.0
func Execute(mlp *MultiLayerNetwork, s *Pattern, options ...int) (r []float64)
Execute a multi layer Perceptron neural network. [mlp:MultiLayerNetwork] multilayer perceptron network pointer, [s:Pattern] input value It returns output values by network
func HeavysideTransfer ¶ added in v0.2.0
func HeavysideTransferDerivate ¶ added in v0.2.0
func HyperbolicTransfer ¶ added in v0.2.0
func HyperbolicTransferDerivate ¶ added in v0.2.0
func MLPTrain ¶ added in v0.2.0
func MLPTrain(mlp *MultiLayerNetwork, patterns []Pattern, mapped []string, epochs int)
MLPTrain train a mlp MultiLayerNetwork with BackPropagation algorithm for assisted learning.
func Predict ¶
func Predict(neuron *NeuronUnit, pattern *Pattern) float64
Predict performs a neuron prediction to passed pattern. It returns a float64 binary predicted value.
func RandomNeuronInit ¶ added in v0.2.0
func RandomNeuronInit(neuron *NeuronUnit, dim int)
RandomNeuronInit initialize neuron weight, bias and learning rate using NormFloat64 random value.
func RawExpectedConversion ¶
RawExpectedConversion converts (string) raw expected values in patterns training / testing sets to float64 values It works on pattern struct (pointer) passed. It doens't returns nothing
func SigmoidalTransfer ¶ added in v0.2.0
func SigmoidalTransferDerivate ¶ added in v0.2.0
func TrainNeuron ¶ added in v0.2.0
func TrainNeuron(neuron *NeuronUnit, patterns []Pattern, epochs int, init int)
TrainNeuron trains a passed neuron with patterns passed, for specified number of epoch. If init is 0, leaves weights unchanged before training. If init is 1, reset weights and bias of neuron before training.
func UpdateWeights ¶
func UpdateWeights(neuron *NeuronUnit, pattern *Pattern) (float64, float64)
UpdateWeights performs update in neuron weights with respect to passed pattern. It returns error of prediction before and after updating weights.
Types ¶
type MultiLayerNetwork ¶ added in v0.2.0
type MultiLayerNetwork struct {
// Lrate represents learning rate of neuron
L_rate float64
// NeuralLayers represents layer of neurons
NeuralLayers []NeuralLayer
// Transfer function
T_func transferFunction
// Transfer function derivative
T_func_d transferFunction
}
func PrepareElmanNet ¶ added in v0.2.0
func PrepareElmanNet(i int, h int, o int, lr float64, tf transferFunction, trd transferFunction) (rnn MultiLayerNetwork)
PrepareElmanNet create a recurrent neUral network neural network. [l:[]int] is an int array with layers neurons number [input, ..., output] [lr:int] is the learning rate of neural network [tr:transferFunction] is a transfer function [tr:transferFunction] the respective transfer function derivative
func PrepareMLPNet ¶ added in v0.2.0
func PrepareMLPNet(l []int, lr float64, tf transferFunction, trd transferFunction) (mlp MultiLayerNetwork)
PrepareMLPNet create a multi layer Perceptron neural network. [l:[]int] is an int array with layers neurons number [input, ..., output] [lr:int] is the learning rate of neural network [tr:transferFunction] is a transfer function [tr:transferFunction] the respective transfer function derivative
type NeuralLayer ¶ added in v0.2.0
type NeuralLayer struct {
// NeuronUnits represents NeuronUnits in layer
NeuronUnits []NeuronUnit
// Lrate represents number of NeuronUnit in layer
Length int
}
Level struct represents a simple NeuronUnits network with a slice of n NeuronUnits.
func PrepareLayer ¶ added in v0.2.0
func PrepareLayer(n int, p int) (l NeuralLayer)
PrepareLayer create a NeuralLayer with n NeuronUnits inside [n:int] is an int that specifies the number of neurons in the NeuralLayer [p:int] is an int that specifies the number of neurons in the previous NeuralLayer It returns a NeuralLayer object
type NeuronUnit ¶ added in v0.2.0
type NeuronUnit struct {
// Weights represents NeuronUnit vector representation
Weights []float64
// Bias represents NeuronUnit natural propensity to spread signal
Bias float64
// Lrate represents learning rate of neuron
Lrate float64
// Value represents desired value when loading input into network in Multi NeuralLayer Perceptron
Value float64
// Delta represents delta error for unit
Delta float64
}
NeuronUnit struct represents a simple NeuronUnit network with a slice of n weights.
type Pattern ¶ added in v0.2.0
type Pattern struct {
// Features that describe the pattern
Features []float64
// Raw (usually string) expected value
SingleRawExpectation string
// Numeric representation of expected value
SingleExpectation float64
// Numeric representation of expected value
MultipleExpectation []float64
}
Pattern struct represents one pattern with dimensions and desired value
func CreateRandomPatternArray ¶ added in v0.2.0
CreateRandomPatternArray load a CSV dataset into an array of Pattern.
Source Files