public static List <DataInstance> copyWithWeight(List <DataInstance> set, double weight)
{
List <DataInstance> output = new List <DataInstance>();
foreach (DataInstance old in set)
{
DataInstance instance = DataInstance.copy($"weighted-copy-of-{old.identifier}", old);
instance.setWeight(weight);
output.Add(instance);
}
return(output);
}
public virtual string classify(DataInstance instance)
{
string value_of_instance = instance.getProperty(this.label);
// Check if we should pass this instance down to another node..
foreach (Node child in this.getNodeChildren())
{
if (value_of_instance == child.value_splitter)
{
return(child.classify(instance));
}
}
// Check if we should pass this instance down to one of our leafs.
foreach (Leaf child in this.getLeafChildren())
{
if (value_of_instance == child.value_splitter)
{
return(child.classifier);
}
}
throw new Exception($"Unknown value {value_of_instance} for {this.label}");
}
public string classify(DataInstance instance)
{
return(this.getRoot().classify(instance));
}
private ObservationSet importCSV(string filepath, bool set_classifiers)
{
var reader = new StreamReader(filepath);
int row = 0;
int classifier_column = -1;
// Loop through CSV lines.
while (!reader.EndOfStream)
{
row++;
string line = reader.ReadLine();
var values = line.Split(this.seperator);
if (row > 2)
{
// This is an instance. The attributes have already been established.
DataInstance addition = new DataInstance(ElementHelper.generateElementId('D', instances.Count));
int column = 0;
// Find the properties of this instance.
foreach (string value in values)
{
// If we do not want to set the classifiers of the instances, we need to check we have not arrived at the
// Final column.
if (set_classifiers || column != classifier_column)
{
string key = "ERR";
if (column == classifier_column)
{
key = this.targetAttribute();
}
else
{
key = column_positions[column];
}
if (value == "")
{
addition.setProperty(key, null);
}
else
{
string value_to_set = value;
if (key != this.target && this.attributes[key] == "continuous")
{
value_to_set = value.Replace(',', this.dec);
value_to_set = value.Replace('.', this.dec);
}
addition.setProperty(key, value_to_set);
}
}
column++;
}
this.instances.Add(addition);
}
else
{
if (row == 1)
{
int column = 0;
// This line describes the attribute names.
foreach (string value in values)
{
// Add it to the columns list.
this.column_positions.Add(value);
this.attributes.Add(value, "UNKNOWN");
column++;
}
}
else if (row == 2)
{
int column = 0;
// This line describes the attribute types
foreach (string value in values)
{
if (value == "classifier")
{
// Classifier column.
// First we remove it from the attributes dictionary, since we do not want the classifier to show up in the attributes.
this.attributes.Remove(this.column_positions[column]);
// Then, we make sure that we know the correct classifier.
this.target = this.column_positions[column];
classifier_column = column;
}
else
{
if (value != "nominal" && value != "continuous")
{
throw new Exception($"Unknown attribute type {value}.");
}
this.attributes[this.column_positions[column]] = value;
}
column++;
}
}
}
}
// All data has been considered, let's return a ObservationSet.
return(new ObservationSet(this.exampleSet(), this.targetAttribute(), this.exampleAttributeTypes()));
}
