public static void writeAboutAttribute(XmlWriter writer, SampleSet.AttributeInfo attribuneInfo)
{
writer.WriteStartElement("Attribute");
writer.WriteAttributeString("Name", attribuneInfo.Name);
if (attribuneInfo.LabelsValues.Count > 0)
{
writer.WriteAttributeString("Type", "Enum");
writer.WriteStartElement("Enum");
writer.WriteAttributeString("Count", XmlConvert.ToString(attribuneInfo.LabelsValues.Count));
for (int i = 0; i < attribuneInfo.LabelsValues.Count; i++)
{
writer.WriteStartElement("Enum");
writer.WriteAttributeString("Value", attribuneInfo.LabelsValues[i]);
writer.WriteEndElement();
}
writer.WriteEndElement();
}
else
{
writer.WriteAttributeString("Type", "Interval");
writer.WriteElementString("Min", XmlConvert.ToString(attribuneInfo.Min));
writer.WriteElementString("Max", XmlConvert.ToString(attribuneInfo.Max));
}
writer.WriteEndElement();
}
public static SampleSet LoadLearnFromUFS(XmlDocument Source)
{ //init
SampleSet temp_set = null;
List <SampleSet.AttributeInfo> input_Attribute = new List <SampleSet.AttributeInfo>();
List <SampleSet.RowSample> data_Row = new List <SampleSet.RowSample>();
SampleSet.AttributeInfo output_Attribute = new SampleSet.AttributeInfo();
string opened_dataset;
//parse_start
XmlNode table_node = Source.DocumentElement.SelectSingleNode("descendant::Table[@Type='Training'] "); //We get learning table
if (table_node == null)
{
throw new System.FormatException("В файле нет таблиц данных");
}
opened_dataset = table_node.Attributes.GetNamedItem("Name").Value;
output_Attribute.Name = table_node.Attributes.GetNamedItem("Output").Value;
XmlNode attrib_node = table_node.SelectSingleNode("Attributes"); //We get atribute's tags
int count_attribs = XmlConvert.ToInt32(attrib_node.Attributes.GetNamedItem("Count").Value);
for (int k = 0; k < count_attribs; k++)
{
SampleSet.AttributeInfo temp_attib = new SampleSet.AttributeInfo();
temp_attib.Name = attrib_node.ChildNodes[k].Attributes.GetNamedItem("Name").Value; // We get one attribute tag
foreach (XmlNode Value in attrib_node.ChildNodes[k].ChildNodes)
{
switch (Value.Name)
{
case "Min": { temp_attib.Min = XmlConvert.ToDouble(Value.InnerXml); break; }
case "Max": { temp_attib.Max = XmlConvert.ToDouble(Value.InnerXml); break; }
case "Enum":
{
foreach (XmlNode tempValue in Value.ChildNodes)
{
temp_attib.LabelsValues.Add(tempValue.Attributes.GetNamedItem("Value").Value);
}
break;
}
}
}
if (temp_attib.LabelsValues.Count == 0)
{
temp_attib.Type = SampleSet.AttributeInfo.TypeAttribute.real;
}
else
{
temp_attib.Type = SampleSet.AttributeInfo.TypeAttribute.nominate;
}
if (temp_attib.Name.Equals(output_Attribute.Name, StringComparison.OrdinalIgnoreCase))
{
output_Attribute = temp_attib;
}
else
{
input_Attribute.Add(temp_attib);
}
}
XmlNode rows_node = table_node.SelectSingleNode("Rows"); //We get data rows
int count_input_attrib = count_attribs - 1;
int count_rows = XmlConvert.ToInt32(rows_node.Attributes.GetNamedItem("Count").Value);
for (int r = 0; r < count_rows; r++)
{
string Classifier_value;
double[] double_value = new double[count_input_attrib];
string[] string_value = new string[count_input_attrib];
for (int a = 0; a < count_input_attrib; a++)
{
XmlNode value = rows_node.ChildNodes[r].SelectSingleNode(XmlConvert.EncodeName(input_Attribute[a].Name));
try { double_value[a] = XmlConvert.ToDouble(value.InnerXml); }
catch
{
string_value[a] = value.InnerText;
}
}
XmlNode outvalue = rows_node.ChildNodes[r].SelectSingleNode(XmlConvert.EncodeName(output_Attribute.Name));
Classifier_value = outvalue.InnerText;
double doublevalue = double.NaN;
if (output_Attribute.Type == SampleSet.AttributeInfo.TypeAttribute.real)
{
doublevalue = XmlConvert.ToDouble(Classifier_value);
Classifier_value = "";
}
SampleSet.RowSample temp_rows = new SampleSet.RowSample(double_value, string_value, doublevalue, Classifier_value);
data_Row.Add(temp_rows);
}
temp_set = new SampleSet(opened_dataset, data_Row, input_Attribute, output_Attribute);
if (temp_set.OutputAttribute.Type == SampleSet.AttributeInfo.TypeAttribute.nominate)
{
temp_set.Type = SampleSet.TypeSampleSet.Classifier;
}
else
{
temp_set.Type = SampleSet.TypeSampleSet.Approximation;
}
GC.Collect();
return(temp_set);
}
