def getModelFromDataframe(df,
max_depth=MAX_DEPTH,
min_samples_leaf=MIN_SAMPLES_LEAF):
X_scaled, y = dataToNumpy(df)
return getModelFromNumpy(X_scaled,
y,
maxDepth=max_depth,
minSamplesLeaf=min_samples_leaf)
def trainAndTestFromDataframes(
trainDf,
testDf,
max_depth=MAX_DEPTH,
min_samples_leaf=MIN_SAMPLES_LEAF): # method to be used in backtester
model = getModelFromDataframe(trainDf,
max_depth=max_depth,
min_samples_leaf=min_samples_leaf)
testX, testY = dataToNumpy(testDf, 'testing df')
predictions = model.predict(testX)
return predictions
def trainAndTestFromDataframes(trainDf, testDf):
model = getModelFromDataframe(trainDf)
testX, testY = dataToNumpy(testDf)
predictions = model.predict(testX)
return predictions
def getModelFromDataframe(df): # method to be used in backtester
X_scaled, y = dataToNumpy(df)
return getModelFromNumpy(X_scaled, y)
def getModelFromCSV(csv='data/postsWithDate.csv'):
df = pd.read_csv(csv)
getModelFromDataframe(df)
def trainAndTestFromDataframes(trainDf, testDf):
model = getModelFromDataframe(trainDf)
testX, testY = dataToNumpy(testDf)
predictions = model.predict(testX)
return predictions
if __name__ == '__main__':
X_scaled, y = dataToNumpy('data/postsWithDate.csv')
clf = getModelFromNumpy(X_scaled, y)
length = X_scaled.shape[0]
# MANUAL TUNING DONE HERE
predictions = clf.predict(X_scaled)
predictions_proba = clf.predict_proba(X_scaled)
print("confusion matrix:")
print(confusion_matrix(y, predictions))
print(f'prediction: {clf.predict_proba([X[140]])}')
zeroCounter = 0
oneCounter = 0
def trainAndTestFromDataframes(
trainDf,
testDf,
max_depth=MAX_DEPTH,
min_samples_leaf=MIN_SAMPLES_LEAF): # method to be used in backtester
model = getModelFromDataframe(trainDf,
max_depth=max_depth,
min_samples_leaf=min_samples_leaf)
testX, testY = dataToNumpy(testDf, 'testing df')
predictions = model.predict(testX)
return predictions
if __name__ == '__main__':
X_scaled, y = dataToNumpy('data/bigOne.csv')
clf = getModelFromNumpy(X_scaled, y)
length = X_scaled.shape[0]
# MANUAL TUNING DONE HERE
predictions = clf.predict(X_scaled)
predictions_proba = clf.predict_proba(X_scaled)
print("confusion matrix:")
print(confusion_matrix(y, predictions))
print(f'prediction: {clf.predict_proba([X_scaled[140]])}')
zeroCounter = 0
oneCounter = 0