public double getDistance(AnalyzedData tuple, List <MinMaxKNN> minMax)
{
double result = 0;
int limit = Math.Min(Attributes.Count, Math.Min(tuple.attributes.Count, minMax.Count));
for (int i = 0; i < limit; i++)
{
result += Math.Pow(Attributes[i].getDistance(tuple.Attributes[i], minMax[i].MinValue, minMax[i].MaxValue), 2);
}
return(Math.Sqrt(result));
}
public void findMostLikelyClass(AnalyzedData analyzedData)
{
string bestClass = Classes[0];
double bestProbability = 0;
foreach (var c in Classes)
{
double r = checkProbabilityForClass(analyzedData, c);
if (r > bestProbability)
{
bestProbability = r;
bestClass = c;
}
}
analyzedData.DataClass = bestClass;
}
public void classifyTuple(AnalyzedData tuple)
{
if (leaf != null)
{
tuple.DataClass = leaf;
return;
}
if (checkCondition(tuple.Attributes[nodeAttribute].getValueAsString()))
{
leftChild.classifyTuple(tuple);
}
else
{
rightChild.classifyTuple(tuple);
}
}
private List <AnalyzedData> findNeighbours(AnalyzedData tuple)
{
List <Tuple <AnalyzedData, double> > dataList = new List <Tuple <AnalyzedData, double> >();
foreach (var v in dataInHyperspace.DataValues)
{
dataList.Add(new Tuple <AnalyzedData, double>(v, tuple.getDistance(v, minMax)));
}
dataList = dataList.OrderBy(x => x.Item2).ToList();
List <AnalyzedData> toReturn = new List <AnalyzedData>();
for (int i = 0; i < k; i++)
{
toReturn.Add(dataList[i].Item1);
}
return(toReturn);
}
public double checkProbabilityForClass(AnalyzedData analyzedData, string analyzedClass)
{
double result = 1;
double classDenominator = 0;
double classNominator = Counters[Counters.Count - 1].returnNumberForOneClass(analyzedClass, analyzedClass);
foreach (var v in Counters[Counters.Count - 1].Attributes)
{
classDenominator += v.Number;
}
for (int i = 0; i < analyzedData.Attributes.Count; i++)
{
double nominator = Counters[i].returnNumberForOneClass(analyzedData.Attributes[i].getValueAsString(), analyzedClass) + 1;
double denominator = classNominator + Counters[i].DiffValues;
result = result * nominator / denominator;
}
result = result * classNominator / classDenominator;
return(result);
}
public DataSet(string path, char splitter, bool hasLabel)
{
Headers = new List <string>();
DataValues = new List <AnalyzedData>();
int rid = 1;
try
{
using (var sr = new StreamReader(path))
{
string headersLine = sr.ReadLine();
var headersNames = headersLine.Split(' ');
for (int i = 0; i < headersNames.Length - 1; i++)
{
Headers.Add(headersNames[i]);
}
string newLine;
while (true)
{
newLine = sr.ReadLine();
if (newLine == null)
{
break;
}
var attributes = newLine.Split(splitter);
AnalyzedData toAdd = new AnalyzedData(rid, attributes, Headers, hasLabel);
DataValues.Add(toAdd);
rid++;
}
}
}
catch (IOException e)
{
Console.WriteLine("The file could not be read:");
Console.WriteLine(e.Message);
}
}
public void setClass(AnalyzedData data)
{
var neighbours = findNeighbours(data);
data.DataClass = doVoting(neighbours);
}
