int main(int argc, char* argv[]){
if (argc > 1)
return file_convert(argv[1],out_file_name(argv[1]));
else
return console_convert();
}
int main(int argc, char* argv[]) {
// first prepare some utility objects
VikStd::TOutFile outfile("log.txt");
VikStd::TOutStd outstd;
VikStd::TOutDouble out2(&outfile, &outstd);
VikStd::TCommon common(&out2);
common.NotifyP("Reading data...");
// load data from input file
VikStd::TDataTable tab;
VikStd::TStr temp_s("test.csv");
tab.LoadFromFileX(temp_s);
VikApriori::TAssocRules ar(&common, &tab);
// determine target attribute
VikStd::TStr TargetName = "A35";
int target = tab.FindAttrWithName(TargetName);
if (target < 0)
throw VikStd::TExc(VikStd::TStr("Incompatible data file and parameter Target. No such attribute: ") + TargetName);
if (tab.Attrs[target].AttrType != VikStd::atDiscr)
throw VikStd::TExc("Incompatible data file and parameter Target. Attribute not discrete.");
ar.SetTargetAttribute(target);
// now setup properties for this algorithm
VikStd::TProperties props;
VikStd::TProp prop;
prop.m_c1 = "OutputFile";
prop.m_c2.m_Str = "res.txt";
prop.m_c2.m_Type = VikStd::VTStr;
props.Add(prop);
prop.m_c1 = "MaxRuleLen";
prop.m_c2.m_Str = 2;
prop.m_c2.m_Type = VikStd::VTInt;
props.Add(prop);
prop.m_c1 = "MinSup";
prop.m_c2.m_Double = 0.05;
prop.m_c2.m_Type = VikStd::VTDouble;
props.Add(prop);
prop.m_c1 = "MinConf";
prop.m_c2.m_Double = 0.7;
prop.m_c2.m_Type = VikStd::VTDouble;
props.Add(prop);
// detect feature subset selection type
prop.m_c1 = "FSSType";
prop.m_c2.m_Type = VikStd::VTInt;
// if ((PercentOfSelectedAttributes >= 1) || (PercentOfSelectedAttributes <=0))
// prop.m_c2.m_Int = FSS_None;
// else
prop.m_c2.m_Int = VikApriori::FSS_OddsRatio;
props.Add(prop);
prop.m_c1 = "PercentOfSelectedAttrs";
prop.m_c2.m_Type = VikStd::VTDouble;
prop.m_c2.m_Double = 0.2;
props.Add(prop);
// detect classifier construction algorithm
// 0 - use all rules
// 1 - use n rules
// 2 - use n rules for each class
prop.m_c1 = "ClassConstrAlgo";
prop.m_c2.m_Type = VikStd::VTInt;
prop.m_c2.m_Int = VikApriori::CC_UseAll;
props.Add(prop);
prop.m_c1 = "N";
prop.m_c2.m_Type = VikStd::VTInt;
prop.m_c2.m_Int = 1;
props.Add(prop);
prop.m_c1 = "DefaultRule";
prop.m_c2.m_Type = VikStd::VTBool;
prop.m_c2.m_Bool = true;
props.Add(prop);
//NormalRunAssocRules(ar, props);
/////////////
/*
out2.Put("Running FOIL...\n");
VikRuleClassifiers::TFoil foil_obj;
foil_obj.m_TargetAttr = target;
foil_obj.Grow(&tab);
foil_obj.Output(&out2, temp_s, &tab);
*/
VikStd::TStr out_file_name("res.txt");
VikClustering::DoClustering(temp_s, out_file_name, 10, 5, out2);
return 0;
}
