void
testStore2(void)
{
int i;
long bO, bF, grossBytes;
bO = stoBytesAlloc - stoBytesFree - stoBytesGc;
printforest = 1;
grossBytes = 0;
forest = (struct tree **)stoAlloc(int0,
FOREST_SIZE*sizeof(struct tree *));
for (i = 0; i < FOREST_SIZE; i++) forest[i] = 0;
randomForest();
for (i = 0; i < FOREST_SIZE; i++) {
long sz = treeSize(forest[i]);
printf("%d: %ld bytes\n", i, sz);
grossBytes += sz;
if (printforest) treePrint(osStdout, forest[i]);
}
for (i = 0; i < FOREST_SIZE; i++) {
treeCheck(forest[i], treeNo);
treeFree(forest[i]);
}
bF = stoBytesAlloc - stoBytesFree - stoBytesGc;
printf("Memory: %lu (owned) %lu (alloc) %ld (net) %ld (gross)\n",
stoBytesOwn, bF - bO, treeBytes, grossBytes);
stoFree(forest);
}
void ClassifyRFSharedCommand::processSharedAndDesignData(vector<SharedRAbundVector*> lookup){
try {
map<string, int> treatmentToIntMap;
map<int, string> intToTreatmentMap;
//vector<string> groups = designMap.getCategory();
for (int i = 0; i < lookup.size(); i++) {
string treatmentName = designMap.get(lookup[i]->getGroup());
treatmentToIntMap[treatmentName] = i;
intToTreatmentMap[i] = treatmentName;
}
int numSamples = lookup.size();
int numFeatures = lookup[0]->getNumBins();
int numRows = numSamples;
int numColumns = numFeatures + 1; // extra one space needed for the treatment/outcome
vector< vector<int> > dataSet(numRows, vector<int>(numColumns, 0));
vector<string> names;
for (int i = 0; i < lookup.size(); i++) {
string sharedGroupName = lookup[i]->getGroup();
names.push_back(sharedGroupName);
string treatmentName = designMap.get(sharedGroupName);
int j = 0;
for (; j < lookup[i]->getNumBins(); j++) {
int otuCount = lookup[i]->get(j);
dataSet[i][j] = otuCount;
}
dataSet[i][j] = treatmentToIntMap[treatmentName];
}
RandomForest randomForest(dataSet, numDecisionTrees, treeSplitCriterion, doPruning, pruneAggressiveness, discardHighErrorTrees, highErrorTreeDiscardThreshold, optimumFeatureSubsetSelectionCriteria, featureStandardDeviationThreshold);
randomForest.populateDecisionTrees();
randomForest.calcForrestErrorRate();
randomForest.printConfusionMatrix(intToTreatmentMap);
map<string, string> variables;
variables["[filename]"] = outputDir + util.getRootName(util.getSimpleName(sharedfile)) + "RF.";
variables["[distance]"] = lookup[0]->getLabel();
string filename = getOutputFileName("summary", variables);
outputNames.push_back(filename); outputTypes["summary"].push_back(filename);
randomForest.calcForrestVariableImportance(filename);
//
map<string, string> variable;
variable["[filename]"] = outputDir + util.getRootName(util.getSimpleName(sharedfile)) + "misclassifications.";
variable["[distance]"] = lookup[0]->getLabel();
string mc_filename = getOutputFileName("summary", variable);
outputNames.push_back(mc_filename); outputTypes["summary"].push_back(mc_filename);
randomForest.getMissclassifications(mc_filename, intToTreatmentMap, names);
//
m->mothurOutEndLine();
}
catch(exception& e) {
m->errorOut(e, "ClassifySharedCommand", "processSharedAndDesignData");
exit(1);
}
}