/// <summary>
/// Menu: Rename column
/// </summary>
void _mnuRenameColumn_Click(object sender, EventArgs e)
{
string newName = FrmInputSingleLine.Show(this._listView.FindForm(), "Rename column", this._clickedColumn.Id, "Enter a new name for this column", this._clickedColumn.OverrideDisplayName);
if (newName != null)
{
this._clickedColumn.OverrideDisplayName = newName;
this._clickedColumn.Header.Text = this._clickedColumn.ToString();
this.SaveColumnUserPreferences();
}
}
private void _btnEditId_Click(object sender, EventArgs e)
{
FrmMsgBox.ShowWarning(this, "Edit ID", "The ID represents the identifier used when the data was first loaded. It must be unique. Changing the ID may have unintended consequences.", FrmMsgBox.EDontShowAgainId.ChangeExperimentalGroupsId);
string newId = FrmInputSingleLine.Show(this, this.Text, "Edit ID", this._group.DisplayName, this._txtId.Text);
if (newId != null)
{
this._txtId.Text = newId;
}
}
void set_peak_names()
{
string header = FrmInputSingleLine.Show(this, this.Text, "Peak names", "Enter the peak names", "{DisplayName}");
if (header != null)
{
ParseElementCollection hc = new ParseElementCollection(header);
foreach (Peak p in this._core.Peaks)
{
p.OverrideDisplayName = hc.ConvertToString(p, this._core);
}
}
}
void find_classifier()
{
Core core = this._core;
UserFlag type1;
UserFlag type2;
ConfigurationStatistic stat = DataSet.ForStatistics(this._core).ShowList(this, null);
if (stat == null)
{
FrmMsgBox.ShowError(this, "No stat with this name");
return;
}
string sign = FrmInputSingleLine.Show(this, "Classifier settings", "Find classifier", "Enter the cutoff, or 0 for for automatic", "0");
double manCutoff;
type1 = DataSet.ForUserFlags(this._core).IncludeMessage("Specify the comment flag signifying the first type").ShowList(this, null);
if (type1 == null)
{
return;
}
type2 = DataSet.ForUserFlags(this._core).IncludeMessage("Specify the comment flag signifying the second type").ShowList(this, null);
if (type2 == null)
{
return;
}
if (!double.TryParse(sign, out manCutoff))
{
return;
}
StringBuilder sb = new StringBuilder();
// TID0 = Full
// TID1..5 = Test/training
// TID6..9 = Bootstrap
for (int tid = 0; tid < 10; tid++)
{
// Get all significances
List <double> sigs = new List <double>(core.Peaks.Select(λ => λ.GetStatistic(stat)));
List <bool> inTrainingSet = new List <bool>(core.Peaks.Count);
int co;
// For the training only include 75%
if (tid >= 1 && tid <= 5)
{
co = (int)(core.Peaks.Count * 0.75d);
}
else
{
co = core.Peaks.Count;
}
for (int n = 0; n < core.Peaks.Count; n++)
{
inTrainingSet.Add(n < co);
}
inTrainingSet.Shuffle();
// For the boot-strap shuffle the sigs
if (tid >= 6)
{
sigs.Shuffle();
}
Tuple <double, int, int, int, int> best = null;
Tuple <double, int, int, int, int> bestTest = null;
double cutoff = 0;
// Find the best variable cutoff
if (manCutoff == 0.0d)
{
for (int n = 0; n < core.Peaks.Count; n++)
{
if (inTrainingSet[n])
{
var success = this.SimpleClassify(sigs[n], type1, type2, sigs, inTrainingSet, true);
if (best == null || success.Item1 > best.Item1)
{
best = success;
bestTest = this.SimpleClassify(sigs[n], type1, type2, sigs, inTrainingSet, false);
cutoff = sigs[n];
}
}
}
}
else
{
best = this.SimpleClassify(manCutoff, type1, type2, sigs, inTrainingSet, true);
bestTest = this.SimpleClassify(manCutoff, type1, type2, sigs, inTrainingSet, false);
cutoff = manCutoff;
}
sb.AppendLine(tid == 0 ? "FULLDATA" : tid <= 5 ? "VALIDATION" : "BOOTSTRAP");
sb.AppendLine();
sb.AppendLine(" " + type1 + " <= " + cutoff + " < " + type2);
sb.AppendLine();
sb.AppendLine(" TRAINING SET (" + co + ")");
sb.AppendLine(" " + type1 + " correct: " + StringHelper.AsFraction(best.Item2, best.Item2 + best.Item3));
sb.AppendLine(" " + type2 + " correct: " + StringHelper.AsFraction(best.Item4, best.Item4 + best.Item5));
sb.AppendLine(" Total correct: " + StringHelper.AsFraction(best.Item2 + best.Item4, best.Item2 + best.Item4 + best.Item3 + best.Item5));
sb.AppendLine(" Variables used: " + StringHelper.AsFraction(best.Item2 + best.Item4 + best.Item3 + best.Item5, core.Peaks.Count));
sb.AppendLine();
if (co != core.Peaks.Count)
{
sb.AppendLine(" TEST SET (" + (core.Peaks.Count - co) + ")");
sb.AppendLine(" " + type1 + " correct: " + StringHelper.AsFraction(bestTest.Item2, bestTest.Item2 + bestTest.Item3));
sb.AppendLine(" " + type2 + " correct: " + StringHelper.AsFraction(bestTest.Item4, bestTest.Item4 + bestTest.Item5));
sb.AppendLine(" Total correct: " + StringHelper.AsFraction(bestTest.Item2 + bestTest.Item4, bestTest.Item2 + bestTest.Item4 + bestTest.Item3 + bestTest.Item5));
sb.AppendLine(" Variables used: " + StringHelper.AsFraction(bestTest.Item2 + bestTest.Item4 + bestTest.Item3 + bestTest.Item5, core.Peaks.Count));
sb.AppendLine();
sb.AppendLine(" SCORE: " + (bestTest.Item1 * 100).ToString("F02"));
}
else
{
sb.AppendLine(" SCORE: " + (best.Item1 * 100).ToString("F02"));
}
sb.AppendLine();
sb.AppendLine("--------------------------------------------------------------------------------");
sb.AppendLine();
}
FrmInputMultiLine.ShowFixed(this, "Find classifier", "Classifier results", "Best value to determine split between variables marked with \"" + type1 + "\" and \"" + type2 + "\" based on their significances", sb.ToString());
}