public void Copy(FileInfo source, FileInfo target, int start, int stop, int size)
{
var inputField = new IInputField[55];
var norm = new DataNormalization {
Report = this, Storage = new NormalizationStorageCSV(target.ToString())
};
for (int i = 0; i < 55; i++)
{
inputField[i] = new InputFieldCSV(true, source.ToString(), i);
norm.AddInputField(inputField[i]);
IOutputField outputField = new OutputFieldDirect(inputField[i]);
norm.AddOutputField(outputField);
}
// load only the part we actually want, i.e. training or eval
var segregator2 = new IndexSampleSegregator(start, stop, size);
norm.AddSegregator(segregator2);
norm.Process();
}
public void Narrow(FileInfo source, FileInfo target, int field, int count)
{
var inputField = new IInputField[55];
var norm = new DataNormalization {
Report = this, Storage = new NormalizationStorageCSV(target.ToString())
};
for (int i = 0; i < 55; i++)
{
inputField[i] = new InputFieldCSV(true, source.ToString(), i);
norm.AddInputField(inputField[i]);
IOutputField outputField = new OutputFieldDirect(inputField[i]);
norm.AddOutputField(outputField);
}
var segregator = new IntegerBalanceSegregator(inputField[field], count);
norm.AddSegregator(segregator);
norm.Process();
Console.WriteLine(@"Samples per tree type:");
Console.WriteLine(segregator.DumpCounts());
}