if (docrossvalidation == 1):
df_scores = cross_validation_mse(names, classifiers, X_train, y_train, 10,
ncores)
plot_cross_validation_mse(names, df_scores, suffix)
if (doRoCLearning == 1):
confusion(mylistvariables, names, classifiers, suffix, X_train, y_train, 5)
precision_recall(mylistvariables, names, classifiers, suffix, X_train,
y_train, 5)
plot_learning_curves(names, classifiers, suffix, X_train, y_train, 100,
12000, 300)
if (dotesting == 1):
filenametest_set_ML = "output/testsample%sMLdecision.pkl" % (suffix)
ntuplename = "fTreeFlagged%s" % (optionClassification)
test_setML = test(names, trainedmodels, X_test, test_set)
test_set.to_pickle(filenametest_set_ML)
if (doBoundary == 1):
mydecisionboundaries = decisionboundaries(names, trainedmodels, suffix,
X_train, y_train)
X_train_2PC, pca = GetPCADataFrameAndPC(X_train, 2)
trainedmodels = fit(names, classifiers, X_train_2PC, y_train)
mydecisionboundaries = decisionboundaries(names, trainedmodels,
suffix + "PCAdecomposition",
X_train_2PC, y_train)
if (doBinarySearch == 1):
namesCV, classifiersCV, param_gridCV, changeparameter = getgridsearchparameters(
optionClassification)
grid_search_models, grid_search_bests = do_gridsearch(
if (activateKerasModels == 1):
classifiersDNN, namesDNN = getclassifiersDNN(len(X_train.columns))
classifiers = classifiers + classifiersDNN
names = names + namesDNN
if (dotraining == 1):
trainedmodels = fit(names, classifiers, X_train, y_train)
savemodels(names, trainedmodels, output, suffix)
if (dotesting == 1):
filenametest_set_ML = output + "/testsample%sMLdecision.pkl" % (suffix)
filenametest_set_ML_root = output + "/testsample%sMLdecision.root" % (
suffix)
ntuplename = getTreeName(optionClassification) + "Tested"
test_setML = test(names, trainedmodels, test_set, mylistvariables,
myvariablesy)
test_setML.to_pickle(filenametest_set_ML)
writeTree(filenametest_set_ML_root, ntuplename, test_setML)
if (doRoCLearning == 1):
# confusion(mylistvariables,names,classifiers,suffix,X_train,y_train,5)
precision_recall(mylistvariables, names, classifiers, suffix, X_train,
y_train, 5, plotdir)
plot_learning_curves(names, classifiers, suffix, plotdir, X_train, y_train,
500, nevents, 4000)
if (doOptimisation == 1):
if not ((classtype == "HFmeson") & (optionClassification == "Ds")):
print("==================ERROR==================")
print(
"Optimisation is not implemented for this classification problem. The code is going to fail"