public NaiveBayes(DataSet dataSet)
{
Headers = new List <string>();
Classes = new List <string>();
Counters = new List <AttributeCounter>();
foreach (var v in dataSet.Headers)
{
Headers.Add(v);
AttributeCounter toAdd = new AttributeCounter();
Counters.Add(toAdd);
}
Headers.Add("Class");
Counters.Add(new AttributeCounter());
foreach (var v in dataSet.DataValues)
{
for (int i = 0; i < v.Attributes.Count; i++)
{
Counters[i].addAttribute(v.Attributes[i].getValueAsString(), v.DataClass);
}
Counters[Counters.Count - 1].addAttribute(v.DataClass, v.DataClass);
}
Counters[Counters.Count - 1].DiffValues = Counters[Counters.Count - 1].Attributes.Count;
foreach (var v in Counters[Counters.Count - 1].Attributes)
{
Classes.Add(v.Name);
}
}
public KNN(DataSet data, int kn)
{
dataInHyperspace = data;
k = kn;
minMax = new List <MinMaxKNN>();
classesCounter = new AttributeCounter();
setMinMaxValuesAndCountClasses();
}
private string doVoting(List <AnalyzedData> data)
{
AttributeCounter counter = new AttributeCounter();
foreach (var v in data)
{
counter.addAttribute(v.DataClass, v.DataClass);
}
List <Tuple <string, double> > votes = new List <Tuple <string, double> >();
foreach (var v in counter.Attributes)
{
double x = (double)v.Number / (double)classesCounter.returnNumberForOneClass(v.Name, v.Name);
Tuple <string, double> toAdd = new Tuple <string, double>(v.Name, x);
votes.Add(toAdd);
}
votes = votes.OrderByDescending(x => x.Item2).ToList();
return(votes[0].Item1);
}
