public SplittedData SplitAndProcess(byte[] buffer)
{
_buffer = buffer;
_splittedData = new SplittedData();
for (_currentPosition = 0; _currentPosition < buffer.Length; _currentPosition++)
{
if (buffer.Length - _currentPosition < 10)
{
break;
}
var currentByte = buffer[_currentPosition];
if (currentByte == 0xA5)
{
_currentPosition += ProcessTriggerFrame();
}
else
{
var nextByte = buffer[_currentPosition + 1];
if (nextByte == 0x68)
{
_currentPosition += ProcessWeightFrame();
}
else
{
break;
}
}
}
return(_splittedData);
}
public void TestInformationGainCalculator_WithGroupsCountsOnly()
{
// Given
var initialDataTable = new DataFrame(
new DataTable()
{
Columns = { new DataColumn("Col1", typeof(string)) },
Rows =
{
new object[] { "A" },
new object[] { "A" },
new object[] { "A" },
new object[] { "B" },
new object[] { "B" },
}
});
var group1 = initialDataTable.GetSubsetByRows(new[] { 0, 1, 3 });
var splittedDataGroup1 = new SplittedData(null, group1);
var group1UniqueValuesCount =
group1.GetColumnVector("Col1")
.Values.GroupBy(val => val, val => val)
.Select(grp => grp.Count())
.ToList();
var group2 = initialDataTable.GetSubsetByRows(new[] { 2, 4 });
var splittedDataGroup2 = new SplittedData(null, group2);
var group2UniqueValuesCount =
group2.GetColumnVector("Col1")
.Values.GroupBy(val => val, val => val)
.Select(grp => grp.Count())
.ToList();
var groupsCounts = new List<IList<int>> { group1UniqueValuesCount, group2UniqueValuesCount };
// When
var entropyFromGroups = informationGainCalculator.CalculateSplitQuality(
1.0,
5,
new List<ISplittedData> { splittedDataGroup1, splittedDataGroup2 },
"Col1");
var entropyFromGroupCounts = informationGainCalculator.CalculateSplitQuality(1.0, 5, groupsCounts);
// Then
Assert.AreEqual(entropyFromGroupCounts, entropyFromGroups);
}
private void SaveData(SplittedData data)
{
data.TriggerData.ForEach(t => _dataRepository.SaveTriggerData(t));
data.WeightData.ForEach(w => _dataRepository.SaveWeigthData(w));
}
