def trainWithHotEncoding(hot_encoded_train_features, hot_encoded_train_labels,
hot_encoded_test_features, hot_encoded_test_labels,
results, algorithms, isTesting):
# TRAIIN RANDOM FOREST
class_weights = class_weight.compute_class_weight(
'balanced', np.unique(hot_encoded_train_labels.flatten()),
hot_encoded_train_labels.flatten())
param_grid = {
'max_features': [
# None,
# "sqrt",
"log2"
],
'n_estimators': [
# 1000,
2000,
# 3000
]
}
if (isTesting):
print("TESTING MODE RF: ONLY TRAINING 1 MODEL")
param_grid = {}
rf_hot_encoding = RandomForest(class_weights=class_weights,
param_grid=param_grid)
rf_hot_encoding.title = "RANDOM FOREST HOT ENCODING TRAIN"
rf_hot_encoding.train(hot_encoded_train_features, hot_encoded_train_labels)
model_name = 'RF-HOT.model'
print("SAVING MODEL: ", model_name)
try:
joblib.dump(rf_hot_encoding.model, model_name)
except Exception as e:
print("Cannot save {} because: \n\n".format(model_name), str(e))
algorithms["RANDOM FOREST HOT ENCODING"] = rf_hot_encoding
rf_hot_encoding.drawCurves(X=hot_encoded_train_features,
y=hot_encoded_train_labels)
rf_hot_encoding.title = "RANDOM FOREST HOT ENCODING TEST"
results["RANDOM FOREST HOT ENCODING"] = rf_hot_encoding.drawCurves(
X=hot_encoded_test_features, y=hot_encoded_test_labels)
return rf_hot_encoding, results, algorithms
def trainWithFrecuencies(tetra_freq_train_features, tetra_freq_train_labels,
tetra_freq_test_features, tetra_freq_test_labels,
results, algorithms, isTesting):
class_weights = class_weight.compute_class_weight(
'balanced', np.unique(tetra_freq_train_labels.flatten()),
tetra_freq_train_labels.flatten())
param_grid = {
'max_features': [
# None,
# "sqrt",
"log2"
],
'n_estimators': [
# 1000,
2000,
# 3000
]
}
if (isTesting):
print("TESTING MODE RF: ONLY TRAINING 1 MODEL")
param_grid = {}
rf = RandomForest(class_weights=class_weights, param_grid=param_grid)
rf.title = "RANDOM FOREST TETRA NUCLEOTIDE FREQUENCY TRAIN"
rf.train(tetra_freq_train_features, tetra_freq_train_labels)
model_name = 'RF-TETRA.model'
print("SAVING MODEL USING JOBLIB: ", model_name)
try:
joblib.dump(rf.model, model_name)
except Exception as e:
print("Cannot save {} because: \n\n".format(model_name), str(e))
algorithms["RANDOM FOREST"] = rf
rf.drawCurves(X=tetra_freq_train_features, y=tetra_freq_train_labels)
rf.title = "RANDOM FOREST TETRA NUCLEOTIDE FREQUENCY TEST"
results["RANDOM FOREST"] = rf.drawCurves(X=tetra_freq_test_features,
y=tetra_freq_test_labels)
return rf, results, algorithms
