NAME
pkregann - regression with artificial neural network (multi-layer perceptron)
SYNOPSIS
FCpkregannF[0] FC-iF[0] input FC-tF[0] training [FC-icF[0] col] [FC-ocF[0] col] FC-oF[0] output [options] [advanced options]
DESCRIPTION
pkregann performs a regression based on an artificial neural network. The regression is trained from the input (FC-icF[0]) and output (FC-ocF[0]) columns in a training text file. Each row in the training file represents one sampling unit. Multi-dimensional input features can be defined with multiple input options (e.g., FC-icF[0] 0 FC-icF[0] 1 FC-icF[0] 2 for three dimensional features).
OPTIONS
- FC-iF[0] filename, FC--inputF[0] filename
-
input ASCII file
- FC-tF[0] filename, FC--trainingF[0] filename
-
training ASCII file (each row represents one sampling unit. Input features should be provided as columns, followed by output)
- FC-oF[0] filename, FC--outputF[0] filename
-
output ASCII file for result
- FC-icF[0] col, FC--inputColsF[0] col
-
input columns (e.g., for three dimensional input data in first three columns use: FC-icF[0] 0 FC-icF[0] 1 FC-icF[0] 2
- FC-ocF[0] col, FC--outputColsF[0] col
-
output columns (e.g., for two dimensional output in columns 3 and 4 (starting from 0) use: FC-ocF[0] 3 FC-ocF[0] 4
- FC-fromF[0] row, FC--fromF[0] row
-
start from this row in training file (start from 0)
- FC-toF[0] row, FC--toF[0] row
-
read until this row in training file (start from 0 or set leave 0 as default to read until end of file)
- FC-cvF[0] size, FC--cvF[0] size
-
n-fold cross validation mode
- FC-nnF[0] number, FC--nneuronF[0] number
-
number of neurons in hidden layers in neural network (multiple hidden layers are set by defining multiple number of neurons: FC-nF[0] 15 FC-nF[0] 1, default is one hidden layer with 5 neurons)
- FC-vF[0] level, FC--verboseF[0] level
-
set to: 0 (results only), 1 (confusion matrix), 2 (debug)
Advanced options
- FC--offsetF[0] value
-
offset value for each spectral band input features: FCrefl[band]=(DN[band]-offset[band])/scale[band]F[0]
- FC--scaleF[0] value
-
scale value for each spectral band input features: FCrefl=(DN[band]-offset[band])/scale[band]F[0] (use FC0F[0] if scale min and max in each band to FC-1.0F[0] and FC1.0F[0])
- FC--connectionF[0] rate
-
connection rate (default: 1.0 for a fully connected network)
- FC-lF[0] rate, FC--learningF[0] rate
-
learning rate (default: 0.7)
- FC--maxitF[0] number
-
number of maximum iterations (epoch) (default: 500)