public static void RunC45()
{
List <DDataRecord> records = LoadSample();
C45 <DDataRecord> algorithm = new C45 <DDataRecord>();
algorithm.UpdateContinuousAttributes(records, "temperature");
algorithm.UpdateContinuousAttributes(records, "humidity");
algorithm.Train(records);
//algorithm.RulePostPrune(records); //post pruning using cross valiation set
Console.WriteLine("C4.5 Tree Built!");
for (int i = 0; i < records.Count; i++)
{
DDataRecord rec = records[i];
Console.WriteLine("rec: ");
string[] feature_names = rec.FindFeatures();
foreach (string feature_name in feature_names)
{
Console.WriteLine(feature_name + " = " + rec[feature_name]);
}
Console.WriteLine("Label: " + rec.Label);
Console.WriteLine("Predicted Label: " + algorithm.Predict(records[i]));
Console.WriteLine();
}
}
public void TestC45()
{
var samples = new List <IList <object> >
{
new List <object> {
1, "Sunny", 85, 85, false, false
},
new List <object> {
2, "Sunny", 80, 90, true, false
},
new List <object> {
3, "Overcast", 83, 78, false, true
},
new List <object> {
4, "Rainy", 70, 96, false, true
},
new List <object> {
5, "Rainy", 68, 80, false, true
},
new List <object> {
6, "Rainy", 65, 70, true, false
},
new List <object> {
7, "Overcast", 64, 65, true, true
},
new List <object> {
8, "Sunny", 72, 95, false, false
},
new List <object> {
9, "Sunny", 69, 70, false, true
},
new List <object> {
10, "Rainy", 75, 80, false, true
},
new List <object> {
11, "Sunny", 75, 70, true, true
},
new List <object> {
12, "Overcast", 73, 90, true, true
},
new List <object> {
13, "Overcast", 81, 75, false, true
},
new List <object> {
14, "Rainy", 71, 80, true, false
},
};
var optionIndexes = new List <int> {
1, 2, 3, 4
};
var targetIndex = 5;
var algo = new C45 <object>();
algo.Process(samples, optionIndexes, targetIndex);
}
private void RodarArvore(int configuracaoArvoreId)
{
var atributos = db.ConfiguracaoAtributo.Where(config => config.ConfiguracaoArvoreID == configuracaoArvoreId).ToList();
string classeMeta = atributos[0].ClasseMeta;
var configuracaoArvore = db.ConfiguracaoArvore.FirstOrDefault(configArvore => configArvore.ID == configuracaoArvoreId);
//Gerar a árvore
List <Atributo> atrs = new List <Atributo>();
foreach (var atr in atributos)
{
atrs.Add(new Atributo {
Nome = atr.Nome, TipoAtributo = atr.Tipo, Legenda = atr.Legenda
});
}
TabelaTemporaria.PreencherAtributos(atrs, configuracaoArvoreId);
C45 c45 = new C45();
if (configuracaoArvore.Poda > 0)
{
c45.Poda = configuracaoArvore.Poda;
}
c45.Calcular(classeMeta, atrs);
ArvoreGerada arvoreGerada = new ArvoreGerada();
arvoreGerada.ConfiguracaoArvoreID = configuracaoArvoreId;
arvoreGerada.ClasseMeta = atributos.FirstOrDefault(x => x.Nome.Equals(classeMeta)).Legenda;
arvoreGerada.JSON = new JavaScriptSerializer().Serialize(c45.arvore.GerarJSON());
arvoreGerada.UsuarioGeracao = "Agendamento";
arvoreGerada.DataGeracao = DateTime.Now;
var xml = new System.Xml.Serialization.XmlSerializer(c45.arvore.GetType());
using (StringWriter textWriter = new StringWriter())
{
xml.Serialize(textWriter, c45.arvore);
arvoreGerada.XML = textWriter.ToString();
}
var doc = new GeradorPMML().Gerar(atrs, classeMeta, c45.arvore, c45.PercentuaisClasseMeta);
xml = new System.Xml.Serialization.XmlSerializer(doc.GetType());
using (StringWriter textWriter = new StringWriter())
{
xml.Serialize(textWriter, doc);
arvoreGerada.XmlPmml = textWriter.ToString();
}
if (atributos.Count(x => x.Tipo == Tipo.Mineração_de_Texto) > 0)
{
arvoreGerada.JsonNuvemPalavras = new JavaScriptSerializer().Serialize(TabelaTemporaria.RetornaNuvemPalavras(atributos.FirstOrDefault(x => x.Tipo == Tipo.Mineração_de_Texto).Nivel));
}
db.ArvoreGerada.Add(arvoreGerada);
db.SaveChanges();
}
public void TestC45()
{
var c45 = new C45();
var c45Parameter = new C45Parameter(1, true, 0.2);
var discreteToContinuous = new DiscreteToContinuous(car);
discreteToContinuous.Convert();
c45.Train(car.GetInstanceList(), c45Parameter);
Assert.AreEqual(29.98, 100 * c45.Test(car.GetInstanceList()).GetErrorRate(), 0.01);
discreteToContinuous = new DiscreteToContinuous(tictactoe);
discreteToContinuous.Convert();
c45.Train(tictactoe.GetInstanceList(), c45Parameter);
Assert.AreEqual(34.66, 100 * c45.Test(tictactoe.GetInstanceList()).GetErrorRate(), 0.01);
}
public void TestC45()
{
var c45 = new C45();
var c45Parameter = new C45Parameter(1, true, 0.2);
var laryToBinary = new LaryToBinary(car);
laryToBinary.Convert();
c45.Train(car.GetInstanceList(), c45Parameter);
Assert.AreEqual(10.76, 100 * c45.Test(car.GetInstanceList()).GetErrorRate(), 0.01);
laryToBinary = new LaryToBinary(tictactoe);
laryToBinary.Convert();
c45.Train(tictactoe.GetInstanceList(), c45Parameter);
Assert.AreEqual(16.08, 100 * c45.Test(tictactoe.GetInstanceList()).GetErrorRate(), 0.01);
laryToBinary = new LaryToBinary(nursery);
laryToBinary.Convert();
c45.Train(nursery.GetInstanceList(), c45Parameter);
Assert.AreEqual(24.95, 100 * c45.Test(nursery.GetInstanceList()).GetErrorRate(), 0.01);
}
public void TestC45()
{
var c45 = new C45();
var c45Parameter = new C45Parameter(1, true, 0.2);
var discreteToIndexed = new DiscreteToIndexed(car);
discreteToIndexed.Convert();
c45.Train(car.GetInstanceList(), c45Parameter);
Assert.AreEqual(10.76, 100 * c45.Test(car.GetInstanceList()).GetErrorRate(), 0.01);
discreteToIndexed = new DiscreteToIndexed(tictactoe);
discreteToIndexed.Convert();
c45.Train(tictactoe.GetInstanceList(), c45Parameter);
Assert.AreEqual(16.08, 100 * c45.Test(tictactoe.GetInstanceList()).GetErrorRate(), 0.01);
discreteToIndexed = new DiscreteToIndexed(nursery);
discreteToIndexed.Convert();
c45.Train(nursery.GetInstanceList(), c45Parameter);
Assert.AreEqual(24.95, 100 * c45.Test(nursery.GetInstanceList()).GetErrorRate(), 0.01);
}
public void TestTrain()
{
var c45 = new C45();
var c45Parameter = new C45Parameter(1, true, 0.2);
c45.Train(iris.GetInstanceList(), c45Parameter);
Assert.AreEqual(4.67, 100 * c45.Test(iris.GetInstanceList()).GetErrorRate(), 0.01);
c45.Train(bupa.GetInstanceList(), c45Parameter);
Assert.AreEqual(42.03, 100 * c45.Test(bupa.GetInstanceList()).GetErrorRate(), 0.01);
c45.Train(dermatology.GetInstanceList(), c45Parameter);
Assert.AreEqual(6.28, 100 * c45.Test(dermatology.GetInstanceList()).GetErrorRate(), 0.01);
c45.Train(car.GetInstanceList(), c45Parameter);
Assert.AreEqual(21.35, 100 * c45.Test(car.GetInstanceList()).GetErrorRate(), 0.01);
c45.Train(tictactoe.GetInstanceList(), c45Parameter);
Assert.AreEqual(19.94, 100 * c45.Test(tictactoe.GetInstanceList()).GetErrorRate(), 0.01);
c45.Train(nursery.GetInstanceList(), c45Parameter);
Assert.AreEqual(29.03, 100 * c45.Test(nursery.GetInstanceList()).GetErrorRate(), 0.01);
}
public ActionResult Index(IList <MineradorRH.Models.ConfiguracaoAtributo> atributos)
{
string classeMeta = atributos[0].ClasseMeta;
if (string.IsNullOrEmpty(classeMeta))
{
ModelState.AddModelError("", "Um atributo deve ser classe meta.");
return(View(atributos));
}
if (atributos.Count(x => x.Tipo == Tipo.Mineração_de_Texto) > 1)
{
ModelState.AddModelError("", "Árvore pode possuir apenas um atributo do tipo Mineração de texto.");
return(View(atributos));
}
if (ModelState.IsValid)
{
foreach (var atributo in atributos.Where(atr => atr.ID > 0))
{
db.Entry(atributo).State = EntityState.Modified;
}
foreach (var atributo in atributos.Where(atr => atr.ID <= 0))
{
db.ConfiguracaoAtributo.Add(atributo);
}
db.SaveChanges();
int arvoreID = atributos[0].ConfiguracaoArvoreID;
var configuracaoArvore = db.ConfiguracaoArvore.FirstOrDefault(configArvore => configArvore.ID == arvoreID);
//Gerar a árvore
List <Atributo> atrs = new List <Atributo>();
foreach (var atr in atributos)
{
atrs.Add(new Atributo {
Nome = atr.Nome, TipoAtributo = atr.Tipo, Legenda = atr.Legenda, Nivel = atr.Nivel
});
}
TabelaTemporaria.PreencherAtributos(atrs, arvoreID);
C45 c45 = new C45();
if (configuracaoArvore.Poda > 0)
{
c45.Poda = configuracaoArvore.Poda;
}
c45.Calcular(classeMeta, atrs);
ArvoreGerada arvoreGerada = new ArvoreGerada();
arvoreGerada.ConfiguracaoArvoreID = arvoreID;
arvoreGerada.ClasseMeta = atributos.FirstOrDefault(x => x.Nome.Equals(classeMeta)).Legenda;
arvoreGerada.JSON = new JavaScriptSerializer().Serialize(c45.arvore.GerarJSON());
arvoreGerada.UsuarioGeracao = User.Identity.Name;
arvoreGerada.DataGeracao = DateTime.Now;
var xml = new System.Xml.Serialization.XmlSerializer(c45.arvore.GetType());
using (StringWriter textWriter = new StringWriter())
{
xml.Serialize(textWriter, c45.arvore);
arvoreGerada.XML = textWriter.ToString();
}
var doc = new GeradorPMML().Gerar(atrs, classeMeta, c45.arvore, c45.PercentuaisClasseMeta);
xml = new System.Xml.Serialization.XmlSerializer(doc.GetType());
using (StringWriter textWriter = new StringWriter())
{
xml.Serialize(textWriter, doc);
arvoreGerada.XmlPmml = textWriter.ToString();
}
if (atributos.Count(x => x.Tipo == Tipo.Mineração_de_Texto) > 0)
{
arvoreGerada.JsonNuvemPalavras = new JavaScriptSerializer().Serialize(TabelaTemporaria.RetornaNuvemPalavras(atributos.FirstOrDefault(x => x.Tipo == Tipo.Mineração_de_Texto).Nivel));
}
db.ArvoreGerada.Add(arvoreGerada);
db.SaveChanges();
return(View("Grafico", arvoreGerada));
}
return(View(atributos));
}
