/**
* {@inheritDoc}
*/
public function iteration($count = null)
{
assert(is_null($count));
$errorCalculation = new ErrorCalculation();
foreach ($this->training as $pair) {
// calculate the error
$output = $this->network->compute($pair->getInput());
for ($currentAdaline = 0; $currentAdaline < $output->size(); ++$currentAdaline) {
$diff = $pair->getIdeal()->getData($currentAdaline) - $output->getData($currentAdaline);
// weights
for ($i = 0; $i <= $this->network->getInputCount(); ++$i) {
$input = 0.0;
if ($i == $this->network->getInputCount()) {
$input = 1.0;
} else {
$input = $pair->getInput()->getData($i);
}
$this->network->addWeight(0, $i, $currentAdaline, $this->learningRate * $diff * $input);
}
}
$errorCalculation->updateError($output->getData(), $pair->getIdeal()->getData(), $pair->getSignificance());
}
// set the global error
$this->setError($errorCalculation->calculate());
}
/**
* @param MLRegression method
* @param MLDataSet data
* @return double
*/
function calculateRegressionError(MLRegression $method, $data)
{
$errorCalculation = new ErrorCalculation();
if ($method instanceof MLContext) {
$method->clearContext();
}
foreach ($data->getData() as $pair) {
$actual = $method->compute($pair->getInput());
$errorCalculation->updateError($actual->getData(), $pair->getIdeal()->getData(), $pair->getSignificance());
}
return $errorCalculation->calculate();
}
/**
* Calculate the error for this neural network.
* The error is calculated
* using root-mean-square(RMS).
*
* @param
* MLDataSet data
* The training set.
* @return double The error percentage.
*/
public function calculateError(MLDataSet $data)
{
$errorCalculation = new ErrorCalculation();
$actual = array();
$pair = BasicMLDataPair::createPair($data->getInputSize(), $data->getIdealSize());
for ($i = 0; $i < $data->getRecordCount(); ++$i) {
$data->getRecord($i, $pair);
$this->compute($pair->getInputArray(), $actual);
$errorCalculation->updateError($actual, $pair->getIdealArray(), $pair->getSignificance());
}
return $errorCalculation->calculate();
}
