public void ChiSquare04()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(103, 0, 5, 6074);
Assert.AreEqual(0.0D, table.ChiSquareValue(), "FFTQ038");
//TODO BUG (chyba hlasena Karbymu)
//Assert.IsTrue(CloseEnough(6, table.ChiSquareValue(), 0.0550538165123782D));
}
public void ChiSquare01()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(21, 1, 71, 141);
Assert.IsTrue(table.ChiSquareValidity(0.05D), "FFTQ032");
//Assert.AreEqual(1.48276669831304e-08D, table.ChiSquareValue());
Assert.IsTrue(table.ChiSquareValue() < 1e-07D, "FFTQ033"); // correct value should be 1.48276595040141e-08
}
public void ContingecyTable01()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(103, 0, 5, 6074);
Assert.AreEqual(table.K, 108, "FFTQ007");
Assert.AreEqual(table.L, 6074, "FFTQ008");
Assert.AreEqual(table.R, 103, "FFTQ009");
Assert.AreEqual(table.S, 6079, "FFTQ010");
Assert.AreEqual(table.N, 6182, "FFTQ011");
Assert.AreEqual(table.A, 103, "FFTQ012");
Assert.AreEqual(table.B, 0, "FFTQ013");
Assert.AreEqual(table.C, 5, "FFTQ014");
Assert.AreEqual(table.D, 6074, "FFTQ015");
}
public void ContingecyTable02()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(103, 0, 5, 6074, 2);
Assert.AreEqual(table.K, (double)108 / 2, "FFTQ016");
Assert.AreEqual(table.L, (double)6074 / 2, "FFTQ017");
Assert.AreEqual(table.R, (double)103 / 2, "FFTQ018");
Assert.AreEqual(table.S, (double)6079 / 2, "FFTQ019");
Assert.AreEqual(table.N, (double)6182 / 2, "FFTQ020");
Assert.AreEqual(table.A, (double)103 / 2, "FFTQ021");
Assert.AreEqual(table.B, (double)0 / 2, "FFTQ022");
Assert.AreEqual(table.C, (double)5 / 2, "FFTQ023");
Assert.AreEqual(table.D, (double)6074 / 2, "FFTQ024");
}
public void CreateFourFoldTable04()
{
int[][] arrayTable =
new int[3][] {
new int[2] { 0, 5 },
new int[2] { 23, 45 },
new int[2] { 23, 45 }
};
FourFoldContingencyTable table = new FourFoldContingencyTable(arrayTable);
Assert.IsNull(table, "FFTQ005: FFTable was created with bad shape of table (not int[2,2])!");
}
public void SetUp()
{
this.table = new FourFoldContingencyTable(30, 1, 23, 45);
}
/// <summary>
/// Returns the <c>a</c> frequency from 4ft-table.
/// Gets the <c>a</c> frequency of the 4f-table (four fold table). (divided by the
/// <see cref="F:Ferda.Modules.Quantifiers.ContingencyTable.denominator"/>)
/// </summary>
/// <result>The <c>a</c> frequency of the 4f-table (four fold table).</result>
public static double BaseCeilValue(FourFoldContingencyTable table)
{
return table.A;
}
/// <summary>
/// Computes the statistical strength value at the specified significance (alpha).
/// </summary>
/// <returns>Statistical strength value at the specified significance (alpha).</returns>
/// <remarks>
/// <para>Computes the numerical solution of the following equation (for variable p):</para>
/// <para><c>Sum[i = 0..a] x! / (i! * (x - i)!) * p^i * (1.0 - p)^(x - i) - alpha = 0.0</c>,
/// where <c>x = (a + b + c)</c>.</para>
/// <para>The solution must be between 0.0 and 1.0 (inclusive).</para>
/// </remarks>
public override double Value(AbstractQuantifierSetting setting, Ice.Current __current)
{
FourFoldContingencyTable table = new FourFoldContingencyTable(setting.firstContingencyTableRows);
return table.DoubleCriticalImplicationValue(Alpha);
}
/// <summary>
/// Returns XML string with all occurences of Active element "4FT hypothesis".
/// </summary>
/// <param name="index">index of data source in FEplugin data sources table</param>
/// <returns>XML string</returns>
public static string getList(int index)
{
string resultString = ""; // result XML string
// loading DTD to resultString
try { resultString = XMLHelper.loadDTD(); }
catch (Exception e)
{
#if (LADENI)
MessageBox.Show("error while loading DTD: " + e.Message);
#endif
return resultString;
}
// root element
resultString += "<active_list>";
string PropName = "Hypotheses"; // name of Property, which contain list of hypotheses
// searching all boxes with type "4FT-task"
IBoxModule[] FFTTaskBoxes = BoxesHelper.ListBoxesWithID(CFEsourcesTab.Sources[index] as CFEsource, "LISpMinerTasks.FFTTask");
// loop over all boxes - getting all hypotheses
string ErrStr = ""; // error report
string matrix_name = "unknown"; // analyzed data matrix name
string db_name="unknown"; // analyzed database name
string task_name = "unknown"; // task name - given with user name of box FFTTast
#region Loop - processing of each 4ft-Task from array FFTTaskBoxes
foreach (IBoxModule box in FFTTaskBoxes)
{
try
{
// searching data source name (database)
IBoxModule[] db_names = BoxesHelper.ListAncestBoxesWithID(box, "DataMiningCommon.Database");
if (db_names.GetLength(0) != 1) // searched more than one data source or neither one
throw new System.Exception("found " + db_names.GetLength(0).ToString() + " databases");
db_name = (db_names[0].GetPropertyOther("DatabaseName") as StringT).stringValue;
// searching data matrix name
IBoxModule[] matrix_names = BoxesHelper.ListAncestBoxesWithID(box, "DataMiningCommon.DataMatrix");
if (matrix_names.GetLength(0) != 1) // searched more than one data source or neither one
throw new System.Exception("found " + matrix_names.GetLength(0).ToString() + " data matrixes");
matrix_name = (matrix_names[0].GetPropertyOther("Name") as StringT).stringValue;
// searching task name
task_name = box.UserName;
// searching the list of all hypotheses in this task
HypothesesT HypT = box.GetPropertyOther(PropName) as HypothesesT;
HypothesisStruct[] HypList = HypT.hypothesesValue.Clone() as HypothesisStruct[];
// records for storing the results
Rec_hyp_4ft rHyp = new Rec_hyp_4ft(); // hypothesis
Rec_ti_cedent rAnt = new Rec_ti_cedent(); // antecedent
Rec_ti_cedent rSuc = new Rec_ti_cedent(); // succedent
Rec_ti_cedent rCon = new Rec_ti_cedent(); // condition
#region Loop - processing of each hypotez jedne krabicky FFTTask
// Loop over all hypotheses
for (int i = 0; i < HypList.GetLength(0); i++)
{
#region element hyp_4ft
rHyp.id = "hyp4ft" + box.ProjectIdentifier.ToString() + "_" + i.ToString();
rHyp.db_name = db_name;
rHyp.matrix_name = matrix_name;
rHyp.task_name = task_name;
rHyp.antecedent = "ant" + rHyp.id;
rHyp.succedent = "suc" + rHyp.id;
rHyp.condition = "con" + rHyp.id;
FourFoldContingencyTable FFT = new FourFoldContingencyTable(HypList[i].quantifierSetting.firstContingencyTableRows);
rHyp.a = FFT.A.ToString();
rHyp.b = FFT.B.ToString();
rHyp.c = FFT.C.ToString();
rHyp.d = FFT.D.ToString();
// values of quantifiers
rHyp.conf = FourFoldContingencyTable.FoundedImplicationValue(FFT).ToString(); // Founded implication (a/a+b)
rHyp.d_conf = FourFoldContingencyTable.DoubleFoundedImplicationValue(FFT).ToString(); // Double Founded implication (a/a+b+c)
rHyp.e_conf = FourFoldContingencyTable.FoundedEquivalenceValue(FFT).ToString(); // Founded Equivalence (a+d)/(a+b+c+d) ??? BUG ve Ferdovi
rHyp.support = FourFoldContingencyTable.BaseCeilValue(FFT).ToString(); //??? Support (a/a+b+c+d)
// helping variable
double avg_diff = FourFoldContingencyTable.AboveAverageImplicationValue(FFT) - 1; // Averafe difference ((a*(a+b+c+d))/((a+b)*(a+c)) -1)
rHyp.avg_diff = avg_diff.ToString();
rHyp.fisher = FFT.FisherValue().ToString();
rHyp.chi_sq = FFT.ChiSquareValue().ToString();
#endregion
#region element ti_cedent (Antecedent)
rAnt.id = "ant" + rHyp.id;
rAnt.type = "Antecedent";
// literals
int litCounter = 0; // counter of literals of this hypotheses
List<Rec_ti_literal> Lit_a = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Antecedent)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_a.Add(l);
}
}
#endregion
#region element ti_cedent (Succedent)
rSuc.id = "suc" + rHyp.id;
rSuc.type = "Succedent";
// literals
List<Rec_ti_literal> Lit_s = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Succedent)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_s.Add(l);
}
}
#endregion
#region element ti_cedent (Condition)
rCon.id = "con" + rHyp.id;
rCon.type = "Condition";
// literals
List<Rec_ti_literal> Lit_c = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Condition)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_c.Add(l);
}
}
#endregion
#region Generating of one hypotheses to XML string
string oneHypString = "";
// generating hypotheses to XML
oneHypString += rHyp.ToXML();
// generating Antecedent to XML
oneHypString += rAnt.ToXML(Lit_a);
// generating Succedent to XML
oneHypString += rSuc.ToXML(Lit_s);
// generating Condition to XML
oneHypString += rCon.ToXML(Lit_c);
resultString += oneHypString;
#endregion
}
#endregion
}
catch (System.Exception e)
{
ErrStr += "Box ProjectIdentifier=" + box.ProjectIdentifier.ToString() + ": " + e.Message + "\n";
}
}
#endregion
// root element
resultString += "</active_list>";
#if (LADENI)
// generating of error message:
if (!String.IsNullOrEmpty(ErrStr))
MessageBox.Show("Pri nacitani hypotez doslo k chybam:\n" + ErrStr, "Chyba", MessageBoxButtons.OK, MessageBoxIcon.Error);
// Kody - storing output to file "XML4fthypExample.xml" in directory
XMLHelper.saveXMLexample(resultString, "../XML/XML4fthypExample.xml");
#endif
return resultString;
}
/// <summary>
/// The 4ft-quantifier chi-square is valid, if the null hypothesis of independence between antecedent and succedent is rejected,
/// thus we say that antecedent and succedent are not independent.
/// </summary>
/// <returns><c>true</c> if the null hypothesis of independence is rejected.</returns>
public override bool Validity(AbstractQuantifierSetting setting, Ice.Current __current)
{
FourFoldContingencyTable table = new FourFoldContingencyTable(setting.firstContingencyTableRows);
return table.ChiSquareValidity(Alpha);
}
/// <summary>
/// Computes the <c>strength</c>.
/// </summary>
/// <returns>The strength defined as <c>a / (a + b + c)</c>.</returns>
/// <remarks>
/// <para>If (a + b + c) = 0, returns 0.</para>
/// </remarks>
public static double DoubleFoundedImplicationValue(FourFoldContingencyTable table)
{
double a = table.A;
double sum = a + table.B + table.C;
return (sum > 0) ? (a / sum) : 0.0D;
}
public void CreateFourFoldTable02()
{
int[][] arrayTable =
new int[2][] {
new int[2] { -30, 1 },
new int[2] { 23, 45 }
};
FourFoldContingencyTable table = new FourFoldContingencyTable(arrayTable);
Assert.IsNull(table, "FFTQ003: FFTable was created with frequency less than 0!");
}
public void ChiSquare02()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(12, 14, 11, 13);
Assert.IsFalse(table.ChiSquareValidity(0.05D), "FFTQ034");
Assert.IsTrue(closeEnough(6, table.ChiSquareValue(), 0.981874933633231), "FFTQ035");
}
// ==================== 4ft-hypoteza ================================
public static string fFE4fthyp(int index)
{
string resultString = ""; // vysledny XML string
// nacteni DTD do resultStringu
try { resultString = XMLHelper.loadDTD(); }
catch (Exception e)
{
MessageBox.Show("Chyba pri nacitani DTD: " + e.Message);
return resultString;
}
// korenovy element
resultString += "<active_list>";
string PropName = "Hypotheses"; // nazev Property, ktera obsahuje seznam hypotez
// nalezeni vsech krabicek 4FT-uloh
IBoxModule[] FFTTaskBoxes = BoxesHelper.ListBoxesWithID(CFEsourcesTab.Sources[index] as CFEsource, "LISpMinerTasks.FFTTask");
// zpracovani kazde krabicky - ziskani z ni vsechny 4ft-hypotezy
string ErrStr = ""; // zaznam o chybach
string matrix_name = ""; // jmeno analyzovane matice
string db_name = ""; // jmeno analyzovane databaze
string task_name = ""; // jmeno ulohy - dano uzivatelskym nazvem krabicky FFTTast
#region Cyklus - zpracovani vsech 4ft-Tasku z pole FFTTaskBoxes
foreach (IBoxModule box in FFTTaskBoxes)
{
try
{
// nalezeni jmena datoveho zdroje (databaze)
IBoxModule[] db_names = BoxesHelper.ListAncestBoxesWithID(box, "DataMiningCommon.Database");
if (db_names.GetLength(0) != 1) // byl nalezen pocet datovych zdroju ruzny od jedne
throw new System.Exception("bylo nalezeno " + db_names.GetLength(0).ToString() + " databazi");
db_name = (db_names[0].GetPropertyOther("DatabaseName") as StringT).stringValue;
// nalezeni jmena datove matice
IBoxModule[] matrix_names = BoxesHelper.ListAncestBoxesWithID(box, "DataMiningCommon.DataMatrix");
if (matrix_names.GetLength(0) != 1) // byl nalezen pocet datovych matic ruzny od jedne
throw new System.Exception("bylo nalezeno " + matrix_names.GetLength(0).ToString() + " datovych matic");
matrix_name = (matrix_names[0].GetPropertyOther("Name") as StringT).stringValue;
// nalezeni jmena ulohy
task_name = box.UserName;
// nalezeni seznamu vsech hypotez v tomto Tasku
HypothesesT HypT = box.GetPropertyOther(PropName) as HypothesesT;
HypothesisStruct[] HypList = HypT.hypothesesValue.Clone() as HypothesisStruct[];
// recordy pro ukladani vysledku
Rec_hyp_4ft rHyp = new Rec_hyp_4ft(); // hypoteza
Rec_ti_cedent rAnt = new Rec_ti_cedent(); // antecedent
Rec_ti_cedent rSuc = new Rec_ti_cedent(); // succedent
Rec_ti_cedent rCon = new Rec_ti_cedent(); // condition
#region Cyklus - zpracovani vsech hypotez jedne krabicky FFTTask
// cyklus pres vsechny hypotezy
for (int i = 0; i < HypList.GetLength(0); i++)
{
#region element hyp_4ft
rHyp.id = "hyp4ft" + box.ProjectIdentifier.ToString() + "_" + i.ToString();
rHyp.db_name = db_name;
rHyp.matrix_name = matrix_name;
rHyp.task_name = task_name;
rHyp.antecedent = "ant" + rHyp.id;
rHyp.succedent = "suc" + rHyp.id;
rHyp.condition = "con" + rHyp.id;
FourFoldContingencyTable FFT = new FourFoldContingencyTable(HypList[i].quantifierSetting.firstContingencyTableRows);
rHyp.a = FFT.A;
rHyp.b = FFT.B;
rHyp.c = FFT.C;
rHyp.d = FFT.D;
// hodnoty kvantifikatoru
rHyp.conf = FourFoldContingencyTable.FoundedImplicationValue(FFT); // Founded implication (a/a+b)
rHyp.d_conf = FourFoldContingencyTable.DoubleFoundedImplicationValue(FFT); // Double Founded implication (a/a+b+c)
rHyp.e_conf = FourFoldContingencyTable.FoundedEquivalenceValue(FFT); // Founded Equivalence (a+d)/(a+b+c+d) ??? BUG ve Ferdovi
rHyp.support = FourFoldContingencyTable.BaseCeilValue(FFT); //??? Support (a/a+b+c+d)
rHyp.avg_diff = FourFoldContingencyTable.AboveAverageImplicationValue(FFT) - 1; // Averafe difference ((a*(a+b+c+d))/((a+b)*(a+c)) -1)
// tyto hodnoty se vypocitavaji (mozna zbytecne?)
rHyp.fisher = FFT.FisherValue();
rHyp.chi_sq = FFT.ChiSquareValue();
#endregion
// TODO - dodelat vypocty, ostatni ciselne polozky (hodnoty nekterych kvantifikatoru atd.)
#region element ti_cedent (Antecedent)
rAnt.id = "ant" + rHyp.id;
rAnt.type = "Antecedent";
// literaly
int litCounter = 0; // pocitadlo literalu teto hypotezy
ArrayList ARLit_a = new ArrayList();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Antecedent)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if(lit.negation)
l.quant = "¬";
l.quant = lit.literalName;
foreach (string s in lit.categoriesNames)
l.value += s;
ARLit_a.Add(l);
}
}
Rec_ti_literal[] ALit_a = (Rec_ti_literal[])ARLit_a.ToArray(typeof(Rec_ti_literal)); // pole literalu daneho cedentu
#endregion
#region element ti_cedent (Succedent)
rSuc.id = "suc" + rHyp.id;
rSuc.type = "Succedent";
// literaly
ArrayList ARLit_s = new ArrayList();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Succedent)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant = lit.literalName;
foreach (string s in lit.categoriesNames)
l.value += s;
ARLit_s.Add(l);
}
}
Rec_ti_literal[] ALit_s = (Rec_ti_literal[])ARLit_s.ToArray(typeof(Rec_ti_literal)); // pole literalu daneho cedentu
#endregion
#region element ti_cedent (Condition)
rCon.id = "con" + rHyp.id;
rCon.type = "Condition";
// literaly
ArrayList ARLit_c = new ArrayList();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Condition)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant = lit.literalName;
foreach (string s in lit.categoriesNames)
l.value += s;
ARLit_s.Add(l);
}
}
Rec_ti_literal[] ALit_c = (Rec_ti_literal[])ARLit_c.ToArray(typeof(Rec_ti_literal)); // pole literalu daneho cedentu
#endregion
#region Vypsani jedne hypotezy do XML stringu
string oneHypString = "";
// vypsani hypotezy do XML
oneHypString += rHyp.ToXML();
// vypsani Antecedentu do XML
oneHypString += rAnt.ToXML(ALit_a);
// vypsani Succedentu do XML
oneHypString += rSuc.ToXML(ALit_s);
// vypsani Condition do XML
oneHypString += rCon.ToXML(ALit_c);
resultString += oneHypString;
#endregion
}
#endregion
}
catch (System.Exception e)
{
ErrStr += "Box ProjectIdentifier=" + box.ProjectIdentifier.ToString() + ": " + e.Message + "\n";
}
}
#endregion
// vypsani pripadne chybove hlasky:
if (!String.IsNullOrEmpty(ErrStr))
MessageBox.Show("Pri nacitani hypotez doslo k chybam:\n" + ErrStr, "Chyba", MessageBoxButtons.OK, MessageBoxIcon.Error);
// korenovy element
resultString += "</active_list>";
// Kody - ulozeni vystupu do souboru "XML4fthypExample.xml" v adresari
XMLHelper.saveXMLexample(resultString, "../XML/XML4fthypExample.xml");
return resultString;
}
/// <summary>
/// Returns <c>true</c> if the simple deviation strength is greater than or equal to the strength parameter.
/// </summary>
/// <returns><c>true</c> iff <c>(a * d) >= 2^k * (b * c)</c>.</returns>
/// <remarks>
/// <para>There are special cases defined explicitly:</para>
/// <para>If both <c>(a * d) = 0</c> and <c>(b * c) = 0</c>, return true if <c>k <= 0</c>.</para>
/// <para>If only <c>(a * d) = 0</c>, return <c>false</c>.</para>
/// <para>If only <c>(b * c) = 0</c>, return <c>true</c>.</para>
/// </remarks>
public override bool Validity(AbstractQuantifierSetting setting, Ice.Current __current)
{
FourFoldContingencyTable table = new FourFoldContingencyTable(setting.firstContingencyTableRows);
return table.SimpleDeviationValidity(K);
}
/// <summary>
/// Returns XML string with all occurences of Active element "SD-4FFT hypothesis".
/// </summary>
/// <param name="index">index of data source in FEplugin data sources table</param>
/// <returns>XML string</returns>
public static string getList(int index)
{
string resultString = ""; // result XML string
// loading DTD to resultString
try { resultString = XMLHelper.loadDTD(); }
catch (Exception e)
{
#if (LADENI)
MessageBox.Show("error while loading DTD: " + e.Message);
#endif
return resultString;
}
// root element
resultString += "<active_list>";
string PropName = "Hypotheses"; // name of property which contain a list of hypotheses
// searching all boxes SD-4FT-uloh
IBoxModule[] TaskBoxes = BoxesHelper.ListBoxesWithID(CFEsourcesTab.Sources[index] as CFEsource, "LISpMinerTasks.SDFFTTask");
// processing of each box - searching all 4ft-hypotheses
string ErrStr = ""; // error report
string matrix_name = "unknown"; // analyzed data matrix name
string db_name="unknown"; // analyzed database name
string task_name = "unknown"; // task name - given with user name of box FFTTast
// creating delegate of quantifier function - "sum of value"
ContingencyTable.QuantifierValue<FourFoldContingencyTable> sum_delegat = new ContingencyTable.QuantifierValue<FourFoldContingencyTable>(FourFoldContingencyTable.GetSumOfValues);
#region Loop - processing of each SD-4ft-Task from array TaskBoxes
foreach (IBoxModule box in TaskBoxes)
{
try
{
// searching data source name (database)
IBoxModule[] db_names = BoxesHelper.ListAncestBoxesWithID(box, "DataMiningCommon.Database");
if (db_names.GetLength(0) != 1) // searched more than one data source or neither one
throw new System.Exception("found " + db_names.GetLength(0).ToString() + " databases");
db_name = (db_names[0].GetPropertyOther("DatabaseName") as StringT).stringValue;
// searching data matrix name
IBoxModule[] matrix_names = BoxesHelper.ListAncestBoxesWithID(box, "DataMiningCommon.DataMatrix");
if (matrix_names.GetLength(0) != 1) // searched more than one data source or neither one
throw new System.Exception("found " + matrix_names.GetLength(0).ToString() + " data matrixes");
matrix_name = (matrix_names[0].GetPropertyOther("Name") as StringT).stringValue;
// searching task name
task_name = box.UserName;
// searching the list of all hypotheses in this task
HypothesesT HypT = box.GetPropertyOther(PropName) as HypothesesT;
HypothesisStruct[] HypList = HypT.hypothesesValue.Clone() as HypothesisStruct[];
// records for storing the results
Rec_hyp_sd4ft rHyp = new Rec_hyp_sd4ft(); // hypothesis
Rec_ti_cedent rAnt = new Rec_ti_cedent(); // antecedent
Rec_ti_cedent rSuc = new Rec_ti_cedent(); // succedent
Rec_ti_cedent rCon = new Rec_ti_cedent(); // condition
Rec_ti_cedent rSet1 = new Rec_ti_cedent(); // set 1
Rec_ti_cedent rSet2 = new Rec_ti_cedent(); // set 2
#region Loop - processing of each hypotheses of one SDFFTTask box
// Loop over all hypotheses
for (int i = 0; i < HypList.GetLength(0); i++)
{
#region element hyp_sd4ft
rHyp.id = "hypsd4ft" + box.ProjectIdentifier.ToString() + "_" + i.ToString();
rHyp.db_name = db_name;
rHyp.matrix_name = matrix_name;
rHyp.task_name = task_name;
rHyp.antecedent = "ant" + rHyp.id;
rHyp.succedent = "suc" + rHyp.id;
rHyp.condition = "con" + rHyp.id;
rHyp.set1 = "set1" + rHyp.id;
rHyp.set2 = "set2" + rHyp.id;
FourFoldContingencyTable FFT1 = new FourFoldContingencyTable(HypList[i].quantifierSetting.firstContingencyTableRows);
FourFoldContingencyTable FFT2 = new FourFoldContingencyTable(HypList[i].quantifierSetting.secondContingencyTableRows);
rHyp.a = FFT1.A.ToString();
rHyp.b = FFT1.B.ToString();
rHyp.c = FFT1.C.ToString();
rHyp.d = FFT1.D.ToString();
rHyp.e = FFT2.A.ToString();
rHyp.f = FFT2.B.ToString();
rHyp.g = FFT2.C.ToString();
rHyp.h = FFT2.D.ToString();
// values of quantifiers - first set
double conf1 = FourFoldContingencyTable.FoundedImplicationValue(FFT1); // Founded implication (a/a+b)
rHyp.conf1 = conf1.ToString();
double d_conf1 = FourFoldContingencyTable.DoubleFoundedImplicationValue(FFT1); // Double Founded implication (a/a+b+c)
rHyp.d_conf1 = d_conf1.ToString();
double e_conf1 = FourFoldContingencyTable.FoundedEquivalenceValue(FFT1); // Founded Equivalence (a+d)/(a+b+c+d) ??? BUG ve Ferdovi
rHyp.e_conf1 = e_conf1.ToString();
rHyp.support1 = FourFoldContingencyTable.BaseCeilValue(FFT1).ToString(); //??? Support (a/a+b+c+d)
// helping variable
double avg_diff1 = FourFoldContingencyTable.AboveAverageImplicationValue(FFT1) - 1; // Averafe difference ((a*(a+b+c+d))/((a+b)*(a+c)) -1)
rHyp.avg_diff1 = avg_diff1.ToString();
rHyp.fisher1 = FFT1.FisherValue().ToString();
rHyp.chi_sq1 = FFT1.ChiSquareValue().ToString();
// values of quantifiers - second set
double conf2 = FourFoldContingencyTable.FoundedImplicationValue(FFT2); // Founded implication (a/a+b)
rHyp.conf2 = conf2.ToString();
double d_conf2 = FourFoldContingencyTable.DoubleFoundedImplicationValue(FFT2); // Double Founded implication (a/a+b+c)
rHyp.d_conf2 = d_conf2.ToString();
double e_conf2 = FourFoldContingencyTable.FoundedEquivalenceValue(FFT2); // Founded Equivalence (a+d)/(a+b+c+d) ??? BUG ve Ferdovi
rHyp.e_conf2 = e_conf2.ToString();
rHyp.support2 = FourFoldContingencyTable.BaseCeilValue(FFT2).ToString(); //??? Support (a/a+b+c+d)
// helping variable
double avg_diff2 = FourFoldContingencyTable.AboveAverageImplicationValue(FFT2) - 1; // Averafe difference ((a*(a+b+c+d))/((a+b)*(a+c)) -1)
rHyp.avg_diff2 = avg_diff2.ToString();
rHyp.fisher2 = FFT2.FisherValue().ToString();
rHyp.chi_sq2 = FFT2.ChiSquareValue().ToString();
// values of quantifiers - diffrent of first and second set
rHyp.dr_sum = FourFoldContingencyTable.Value<FourFoldContingencyTable>(sum_delegat, FFT1, FFT2, OperationModeEnum.DifferencesOfRelativeFrequencies).ToString();
double df_conf = FourFoldContingencyTable.Combine(conf1, conf2, OperationModeEnum.DifferenceOfQuantifierValues);
rHyp.df_conf = df_conf.ToString();
double df_dfui = FourFoldContingencyTable.Combine(d_conf1, d_conf2, OperationModeEnum.DifferenceOfQuantifierValues);
rHyp.df_dfui = df_dfui.ToString();
double df_fue = FourFoldContingencyTable.Combine(e_conf1, e_conf2, OperationModeEnum.DifferenceOfQuantifierValues);
rHyp.df_fue = df_fue.ToString();
double df_avg = FourFoldContingencyTable.Combine(avg_diff1, avg_diff2, OperationModeEnum.DifferenceOfQuantifierValues);
rHyp.df_avg = df_avg.ToString();
#endregion
#region element ti_cedent (Antecedent)
rAnt.id = rHyp.antecedent;
rAnt.type = "Antecedent";
// literals
int litCounter = 0; // counter of literals of this hypotheses
List<Rec_ti_literal> Lit_a = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Antecedent)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_a.Add(l);
}
}
#endregion
#region element ti_cedent (Succedent)
rSuc.id = rHyp.succedent;
rSuc.type = "Succedent";
// literals
List<Rec_ti_literal> Lit_s = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Succedent)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_s.Add(l);
}
}
#endregion
#region element ti_cedent (Condition)
rCon.id = rHyp.condition;
rCon.type = "Condition";
// literals
List<Rec_ti_literal> Lit_c = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.Condition)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_c.Add(l);
}
}
#endregion
#region element ti_cedent (First Set)
rSet1.id = rHyp.set1;
rSet1.type = "First Set";
// literals
List<Rec_ti_literal> Lit_s1 = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.FirstSet)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_s1.Add(l);
}
}
#endregion
#region element ti_cedent (Second Set)
rSet2.id = rHyp.set2;
rSet2.type = "Second Set";
// literals
List<Rec_ti_literal> Lit_s2 = new List<Rec_ti_literal>();
foreach (BooleanLiteralStruct lit in HypList[i].booleanLiterals)
{
if (lit.cedentType == CedentEnum.SecondSet)
{
Rec_ti_literal l = new Rec_ti_literal();
l.id = "tiLit" + rHyp.id + "_" + litCounter.ToString();
litCounter++;
if (lit.negation)
l.quant = "¬";
l.quant += lit.literalName;
int counter = 0;
foreach (string s in lit.categoriesNames)
{
if (counter > 0)
l.value += ",";
l.value += s;
counter++;
}
Lit_s2.Add(l);
}
}
#endregion
#region Generating of one hypotheses to XML string
string oneHypString = "";
// generating hypotheses to XML
oneHypString += rHyp.ToXML();
// generating Antecedent to XML
oneHypString += rAnt.ToXML(Lit_a);
// generating Succedent to XML
oneHypString += rSuc.ToXML(Lit_s);
// generating Condition to XML
oneHypString += rCon.ToXML(Lit_c);
// generating First Set to XML
oneHypString += rSet1.ToXML(Lit_s1);
// generating Second Set to XML
oneHypString += rSet2.ToXML(Lit_s2);
resultString += oneHypString;
#endregion
}
#endregion
}
catch (System.Exception e)
{
ErrStr += "Box ProjectIdentifier=" + box.ProjectIdentifier.ToString() + ": " + e.Message + "\n";
}
}
#endregion
// root element
resultString += "</active_list>";
#if (LADENI)
// generating of error message:
if (!String.IsNullOrEmpty(ErrStr)) // LADICI
MessageBox.Show("Pri nacitani SD4FT hypotheses doslo k chybam:\n" + ErrStr, "Chyba", MessageBoxButtons.OK, MessageBoxIcon.Error);
// LADICI - Kody - storing output to file "XMLsd4fthypExample.xml" in directory
XMLHelper.saveXMLexample(resultString, "../XML/XMLsd4fthypExample.xml");
#endif
return resultString;
}
/// <summary>
/// Computes the <c>confidence</c>.
/// </summary>
/// <returns>The confidence defined as <c>a / (a + b)</c>.</returns>
/// <remarks>
/// <para>If (a + b) = 0, returns 0.</para>
/// </remarks>
public static double FoundedImplicationValue(FourFoldContingencyTable table)
{
double r = table.R;
return (r > 0) ? (table.A) / r : 0.0D;
}
/// <summary>
/// Computes the <c>strength</c> of founded equivalence.
/// </summary>
/// <returns>The strength defined as <c>(a + d) / (a + b + c + d)</c>.</returns>
/// <remarks>
/// <para>If (a + b + c + d) = 0, returns 0.</para>
/// </remarks>
public static double FoundedEquivalenceValue(FourFoldContingencyTable table)
{
double n = table.N;
return (n > 0) ? (table.A) / n : 0.0D;
}
public void ChiSquare03()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(252, 143, 847, 602);
Assert.IsFalse(table.ChiSquareValidity(0.05D), "FFTQ036");
Assert.IsTrue(closeEnough(6, table.ChiSquareValue(), 0.0550538165123782D), "FFTQ037");
}
public void CreateFourFoldTable05()
{
long[,] arrayTable = new long[2, 2] {
{ 30, 1 },
{ 23, 45 }
};
FourFoldContingencyTable table = new FourFoldContingencyTable(arrayTable);
Assert.IsNotNull(table, "FFTQ006: FFTable wasn`t created!");
}
public void E01()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(103, 0, 5, 6074);
Assert.AreEqual(0.999191200258816D, table.EValue(), "FFTQ041");
Assert.IsTrue(table.EValidity(0.5D), "FFTQ042");
Assert.IsFalse(table.EValidity(0.999999D), "FFTQ043");
}
public void Fischer01()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(103, 0, 5, 6074);
Assert.AreEqual(0.0f, (float)table.FisherValue(), "FFTQ039");
Assert.AreEqual(8.46676800899516e-219D, table.FisherValue(), "FFTQ040");
}
public void CreateFourFoldTable00()
{
int[][] arrayTable = new int[2][] {
new int[2] { 30, 1 },
new int[2] { 23, 45 }
};
FourFoldContingencyTable table = new FourFoldContingencyTable(arrayTable);
Assert.IsNotNull(table, "FFTQ001: FFTable wasn`t created!");
}
public void FoundedImplication02()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(103, 0, 5, 6074);
Assert.AreEqual(1.0D, FourFoldContingencyTable.FoundedImplicationValue(table), "FFTQ026");
}
/// <summary>
/// Returns value of expression (<c>a</c> + <c>d</c>) / (<c>a</c> + <c>b</c> + <c>c</c> + <c>d</c>) from 4ft-table.
/// </summary>
/// <returns>The value of expression (<c>a</c> + <c>d</c>) / (<c>a</c> + <c>b</c> + <c>c</c> + <c>d</c>) 4ft-table.</returns>
public override double Value(AbstractQuantifierSetting setting, Ice.Current __current)
{
FourFoldContingencyTable table = new FourFoldContingencyTable(setting.firstContingencyTableRows);
return table.EValue();
}
public void LowerCriticalImplication01()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(10, 2, 0, 0);
//FourFTQuantifiers.LowerCriticalImplication quantifier = new RelMiner.FourFTQuantifiers.LowerCriticalImplication(0.8f, 0.05f);
Assert.IsFalse(table.CriticalImplicationValidity(0.8D, 0.05D, CoreRelationEnum.LessThanOrEqualCore), "FFTQ028");
Assert.AreEqual(0.561894564884689D, table.CriticalImplicationValue(0.05D), "FFTQ029");
//TODO I`m not sure
}
/// <summary>
/// Returns <c>true</c> if the statistical strength value is greater than or equal to the p parameter with the specified statistical significance (alpha).
/// </summary>
/// <returns><c>true</c> if if the statistical strength value is greater than or equal to the p parameter with the specified statistical significance (alpha).</returns>
/// <remarks>
/// <para>It computes the following condition:</para>
/// <para>Sum[i = 0..a] x! / (i! * (x - i)!) * p^i * (1 - p)^(x - i) is in <c>relation</c> to alpha,
/// where x = (a + b + c).</para>
/// </remarks>
public override bool Validity(AbstractQuantifierSetting setting, Ice.Current __current)
{
FourFoldContingencyTable table = new FourFoldContingencyTable(setting.firstContingencyTableRows);
return table.DoubleCriticalImplicationValidity(P, Alpha, Relation);
}
public void LowerCriticalImplication02()
{
FourFoldContingencyTable table = new FourFoldContingencyTable(100, 1, 0, 0);
Assert.IsTrue(table.CriticalImplicationValidity(0.8D, 0.05D, CoreRelationEnum.LessThanOrEqualCore), "FFTQ030");
Assert.AreEqual(0.953892650075111D, table.CriticalImplicationValue(0.05D), "FFTQ031");
//TODO I`m not sure
}
/// <summary>
/// Computes the above average strength value.
/// </summary>
/// <returns>Above average strength value defined as <c>(a / (a + b)) * ((a + b + c + d) / (a + c))</c> if <c>a > 0</c>; otherwise it returns zero.</returns>
/// <remarks>
/// <para>If a = 0, returns 0.</para>
/// </remarks>
public override double Value(AbstractQuantifierSetting setting, Ice.Current __current)
{
FourFoldContingencyTable table = new FourFoldContingencyTable(setting.firstContingencyTableRows);
return FourFoldContingencyTable.AboveAverageImplicationValue(table);
}
/// <summary>
/// Computes the above average strength result.
/// </summary>
/// <returns>Above average strength result defined as <c>(a / (a + b)) * ((a + b + c + d) / (a + c))</c> if <c>a > 0</c>; otherwise it returns zero.</returns>
/// <remarks>
/// <para>If a = 0, returns 0.</para>
/// </remarks>
public static double AboveAverageImplicationValue(FourFoldContingencyTable table)
{
return (table.A > 0) ? ((table.A * table.N) / (table.R * table.K)) : 0.0D;
}
