ID3Learner LoadDecisionTree(List <Record> trainingSet, ReferenceTable table, int depth)
{
int[][] inputs;
int[] outputs;
var codebook = BuildCodebook(trainingSet, table, out inputs, out outputs);
var attributes = new DecisionVariable[table.Columns.Length - 1];
for (int i = 0; i < attributes.Length; i++)
{
attributes[i] = new DecisionVariable(table.Columns[i], table.GetValues(i).Length);
}
int classCount = 2;
DecisionTree tree = new DecisionTree(attributes, classCount);
ID3LearningEx id3Learning = new ID3LearningEx(tree)
{
MaxHeight = depth
};
id3Learning.Run(inputs, outputs);
return(new ID3Learner(this, tree, codebook, table.Columns.Last()));
}
public List <Record> DiscretizeDataset(List <Record> dataset, ReferenceTable table)
{
int[] row = { 0 };
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
Discretize(table.HasUnknowns(row[0]), rec => double.Parse(rec[row[0]]), (rec, val) => rec[row[0]] = val, table.GetRanges(row[0]), table.GetValues(row[0]), dataset); row[0]++;
return(dataset);
}
Codification BuildCodebook(List <Record> trainingSet, ReferenceTable table, out int[][] inputs, out int[] outputs)
{
DataTable data = new DataTable("Diabetes dataset");
data.Columns.AddRange(Array.ConvertAll(table.Columns, x => new DataColumn(x)));
trainingSet.ForEach(each => data.Rows.Add(each.Values));
Codification codebook = new Codification(data);
DataTable symbols = codebook.Apply(data);
inputs = symbols.ToArray <int>(ExcludeLast(table.Columns));
outputs = symbols.ToArray <int>(table.Columns.Last());
return(codebook);
}
ReferenceTable BuildDataSets(out List <Record> trainingSet, out List <Record> testSet)
{
const string training = @"..\..\Resources\3.1\diabetes_train.txt";
const string test = @"..\..\Resources\3.1\diabetes_test.txt";
ReferenceTable table = new ReferenceTable();
var parser = new RecordParser();
trainingSet = parser.ParseRecords(File.ReadAllLines(training));
testSet = parser.ParseRecords(File.ReadAllLines(test));
parser.DiscretizeDataset(trainingSet, table);
parser.DiscretizeDataset(testSet, table);
return(table);
}
void RunSingleTest(List <Record> trainingSet, ReferenceTable table, List <Record> testSet, int ensembleCount, int depth)
{
Console.WriteLine($"\n[{GetTimeStamp()}] Runing test with Ensemble = {ensembleCount} and MaxDepth = {depth}\n");
Ensemble ensemble = new Ensemble();
Random picker = new Random();
for (int i = 0; i < ensembleCount; i++)
{
ID3Learner learner = LoadDecisionTree(SampledData(trainingSet, picker), table, depth);
ensemble.AddVoter(learner.Predict);
}
ConfusionMatrix testResults = RunTest(ensemble.Test, testSet);
PrintResults(testResults);
}
