fann_scale_input

(PECL fann >= 1.0.0)

fann_scale_input — Scale data in input vector before feed it to ann based on previously calculated parameters

Описание

bool fann_scale_input ( resource $ann , array $input_vector )

Scale data in input vector before feed it to ann based on previously calculated parameters.

Список параметров

ann: Ресурс (resource) нейронной сети.
input_vector: Input vector that will be scaled

Возвращаемые значения

Возвращает TRUE в случае успешного выполнения, или FALSE в противном случае.

Смотрите также

fann_descale_input() - Scale data in input vector after get it from ann based on previously calculated parameters
fann_scale_output() - Scale data in output vector before feed it to ann based on previously calculated parameters

Коментарии

Oct 10

Автор: Nolife


Please note -> ALLfann  scaling related functions are not functional.

They are implemented wrong so the scaling is calculated within the library but it's not referenced back to the input variables.

2017-10-10 07:55:30

http://php5.kiev.ua/manual/ru/function.fann-scale-input.html

Nov 27

Автор: saakyanalexandr at gmail dot com


fann_scale_input and fann_scale_output return not bool value. This function return scaling vector.



Example

$r = fann_scale_input($ann, $input);

$output = fann_run($ann, $input);

$s = fann_scale_output ( $ann, $output);



$r and $s is array

2019-11-27 22:38:15

http://php5.kiev.ua/manual/ru/function.fann-scale-input.html

May 29

Автор: geekgirl dot joy at gmail dot com


<?php



// This example will use the XOR dataset with negative one represented 

// as zero and one represented as one-hundred and demonstrate how to

// scale those values so that FANN can understand them and then how 

// to de-scale the value FANN returns so that you can understand them.



// Scaling allows you to take raw data numbers like -1234.975 or 4502012 

// in your dataset and convert them into an input/output range that

// your neural network can understand. 



// De-scaling lets you take the scaled data and convert it back into 

// the original range.



// scale_test.data

// Note the values are "raw" or un-scaled.

/*

4 2 1

0 0

0

0 100

100

100 0

100

100 100

0

*/



////////////////////

// Configure ANN  //

////////////////////



// New ANN

$ann = fann_create_standard_array(3, [2,3,1]);



// Set activation functions

fann_set_activation_function_hidden($ann, FANN_SIGMOID_SYMMETRIC);

fann_set_activation_function_output($ann, FANN_SIGMOID_SYMMETRIC);



// Read raw (un-scaled) training data from file

$train_data = fann_read_train_from_file("scale_test.data");



// Scale the data range to -1 to 1

fann_set_input_scaling_params($ann , $train_data, -1, 1);

fann_set_output_scaling_params($ann , $train_data, -1, 1);



///////////

// Train //

///////////



// Presumably you would train here (uncomment to perform training)...



// fann_train_on_data($ann, $train_data, 100, 10, 0.01);



// But it's not needed to test the scaling because the training file 

// in this case is just used to compute/derive the scale range. 

// However, doing the training will improve the answer the ANN gives

// in correlation to the training data.



//////////

// Test //

//////////



$raw_input = array(0, 100); // test XOR (0,100) input

$scaled_input = fann_scale_input ($ann , $raw_input); // scaled XOR (-1,1) input

$descaled_input = fann_descale_input ($ann , $scaled_input); // de-scaled XOR (0,100) input

$raw_output = fann_run($ann, $scaled_input); // get the answer/output from the ANN

$output_descale = fann_descale_output($ann, $raw_output); // de-scale the output 



////////////////////

// Report Results //

////////////////////

echo 'The raw_input:' . PHP_EOL;

var_dump($raw_input); 



echo 'The raw_input Scaled then De-Scaled (values are unchanged/correct):' . PHP_EOL;

var_dump($descaled_input); 



echo 'The Scaled input:' . PHP_EOL;

var_dump($scaled_input); 



echo "The raw_output of the ANN (Scaled input):" . PHP_EOL;

var_dump($raw_output);

 

echo 'The De-Scaled output:' . PHP_EOL;

var_dump($output_descale); 

 

 

////////////////////

// Example Output //

////////////////////



 /*

The raw_input:

array(2) {

  [0]=>

  float(0)

  [1]=>

  float(100)

}

The raw_input Scaled then De-Scaled (values are unchanged/correct):

array(2) {

  [0]=>

  float(0)

  [1]=>

  float(100)

}

The Scaled input:

array(2) {

  [0]=>

  float(-1)

  [1]=>

  float(1)

}

The raw_output of the ANN (Scaled input):

array(1) {

  [0]=>

  float(1)

}

The De-Scaled output:

array(1) {

  [0]=>

  float(100)

}

*/

2021-05-29 21:27:05

http://php5.kiev.ua/manual/ru/function.fann-scale-input.html

May 29

Автор: geekgirl dot joy at gmail dot com


<?php



// This example will use the XOR dataset with negative one represented 

// as zero and one represented as one-hundred and demonstrate how to

// scale those values so that FANN can understand them and then how 

// to de-scale the value FANN returns so that you can understand them.



// Scaling allows you to take raw data numbers like -1234.975 or 4502012 

// in your dataset and convert them into an input/output range that

// your neural network can understand. 



// De-scaling lets you take the scaled data and convert it back into 

// the original range.



// scale_test.data

// Note the values are "raw" or un-scaled.

/*

4 2 1

0 0

0

0 100

100

100 0

100

100 100

0

*/



////////////////////

// Configure ANN  //

////////////////////



// New ANN

$ann = fann_create_standard_array(3, [2,3,1]);



// Set activation functions

fann_set_activation_function_hidden($ann, FANN_SIGMOID_SYMMETRIC);

fann_set_activation_function_output($ann, FANN_SIGMOID_SYMMETRIC);



// Read raw (un-scaled) training data from file

$train_data = fann_read_train_from_file("scale_test.data");



// Scale the data range to -1 to 1

fann_set_input_scaling_params($ann , $train_data, -1, 1);

fann_set_output_scaling_params($ann , $train_data, -1, 1);



///////////

// Train //

///////////



// Presumably you would train here (uncomment to perform training)...



// fann_train_on_data($ann, $train_data, 100, 10, 0.01);



// But it's not needed to test the scaling because the training file 

// in this case is just used to compute/derive the scale range. 

// However, doing the training will improve the answer the ANN gives

// in correlation to the training data.



//////////

// Test //

//////////



$raw_input = array(0, 100); // test XOR (0,100) input

$scaled_input = fann_scale_input ($ann , $raw_input); // scaled XOR (-1,1) input

$descaled_input = fann_descale_input ($ann , $scaled_input); // de-scaled XOR (0,100) input

$raw_output = fann_run($ann, $scaled_input); // get the answer/output from the ANN

$output_descale = fann_descale_output($ann, $raw_output); // de-scale the output 



////////////////////

// Report Results //

////////////////////

echo 'The raw_input:' . PHP_EOL;

var_dump($raw_input); 



echo 'The raw_input Scaled then De-Scaled (values are unchanged/correct):' . PHP_EOL;

var_dump($descaled_input); 



echo 'The Scaled input:' . PHP_EOL;

var_dump($scaled_input); 



echo "The raw_output of the ANN (Scaled input):" . PHP_EOL;

var_dump($raw_output);

 

echo 'The De-Scaled output:' . PHP_EOL;

var_dump($output_descale); 

 

 

////////////////////

// Example Output //

////////////////////



 /*

The raw_input:

array(2) {

  [0]=>

  float(0)

  [1]=>

  float(100)

}

The raw_input Scaled then De-Scaled (values are unchanged/correct):

array(2) {

  [0]=>

  float(0)

  [1]=>

  float(100)

}

The Scaled input:

array(2) {

  [0]=>

  float(-1)

  [1]=>

  float(1)

}

The raw_output of the ANN (Scaled input):

array(1) {

  [0]=>

  float(1)

}

The De-Scaled output:

array(1) {

  [0]=>

  float(100)

}

*/