def train():
if request.method == 'POST' and 'inputFiles' in request.files:
file = request.files['inputFiles']
filename = secure_filename(file.filename)
data_reload = FileContents(name=filename)
db.create_all()
db.session.add(data_reload)
db.session.commit()
data_reloaded = FileContents.query.all()
# os.path.join is used so that paths work in every operating system
file.save(os.path.join(folder, filename))
# Data review
new_data = pd.read_csv(os.path.join(folder, str(data_reloaded[-1])))
new_dataplot = new_data.head(10)
new_data_info = new_data.info
new_data_size = new_data.size
new_data_shape = new_data.shape
dropdown_list = list(new_data.columns)
flash('Just a moment, app is thinking!')
if str(data_reloaded[-1]).split('.')[-1] != 'csv':
# Forbidden, No Access
abort(403)
return render_template('train.html',
new_dataplot=new_dataplot,
data_reload=data_reload,
new_data_info=new_data.info,
new_data_size=new_data_size,
new_data_shape=new_data_shape,
dropdown_list=dropdown_list)
def plot_2():
if request.method == 'POST' and 'inputFiles' in request.files:
file = request.files['inputFiles']
filename = secure_filename(file.filename)
data_reload = FileContents(name=filename)
db.create_all()
db.session.add(data_reload)
db.session.commit()
data_reload = FileContents.query.all()
file.save(os.path.join(folder, filename))
new_data = pd.read_csv(os.path.join(folder, str(data_reload[-1])))
dropdown_list = list(new_data.columns)
return render_template('plot_2.html', dropdown_list=dropdown_list)
def fit():
# X and Y vars
y_var_select = request.form.get('y_var')
multiselect = request.form.getlist('x_vars')
# Type of forecast
pred_type_select = request.form.get('rd_pred_type')
# commit the prediction type
pred_type_selected = ModelType(pred_type_select)
db.create_all()
db.session.add(pred_type_selected)
# commit the X and Y vars
xy_selection = ListXY(y_var_select, multiselect)
db.create_all()
db.session.add(xy_selection)
db.session.commit()
# testing - DELETE AFTEWARDS
ListXY.query.all()
# select vars
data_reloaded = FileContents.query.all()
new_data = pd.read_csv(os.path.join(folder, str(data_reloaded[-1])))
new_data = new_data.dropna() # deletes Na and NaN
X = new_data[multiselect]
Y = new_data[y_var_select]
if request.form.get('rd_pred_type') == "Classification":
# Step 1: Refactor columns with text to integer and remove NAs
X = factorise_data(X)
# prepare models
seed = 7
models = []
models.append(('RandomForestClassifier', RandomForestClassifier()))
models.append(
('GradientBoostingClassifier', GradientBoostingClassifier()))
models.append(('LogisticRegression', LogisticRegression()))
models.append(
('LinearDiscriminantAnalysis', LinearDiscriminantAnalysis()))
models.append(('KNeighborsClassifier', KNeighborsClassifier()))
models.append(('GaussianNB', GaussianNB()))
models.append(('SVC', SVC()))
# evaluate each model in turn
results = []
names = []
allmodels = []
scoring = 'accuracy'
for name, model in models:
kfold = model_selection.KFold(n_splits=10, random_state=seed)
cv_results = model_selection.cross_val_score(model,
X,
Y,
cv=kfold,
scoring=scoring)
results.append(cv_results)
names.append(name)
msg = "%s - %f | %f" % (name, cv_results.mean(), cv_results.std())
allmodels.append(msg)
model_results = results
model_names = names
if request.form.get('rd_pred_type') == "Regression":
# Step 1: Refactor columns with text to integer and remove NAs
X = factorise_data(X)
# prepare models
models = []
models.append(
('RandomForestRegressor', RandomForestRegressor(n_estimators=200)))
models.append(('GradientBoostingRegressor',
GradientBoostingRegressor(n_estimators=200)))
models.append(('Ridge', Ridge()))
models.append(('ElasticNet', ElasticNet()))
models.append(('Lasso', Lasso()))
models.append(('SVR', SVR()))
# evaluate each model in turn
results = []
names = []
allmodels = []
for name, model in models:
X_train, X_test, y_train, y_test = train_test_split(X,
Y,
test_size=0.3,
random_state=7)
# standard scaler #standardises the feature variables
sc = StandardScaler()
X_train = sc.fit_transform(X_train)
X_test = sc.transform(X_test)
model_to_fit = model
model_to_fit.fit(X_train, y_train)
predictions = model_to_fit.predict(X_test)
mse = mean_squared_error(y_test, predictions)
results.append(mse)
names.append(name)
msg = "%s - %.2f | %s" % (name, (np.sqrt(mse)), "-")
allmodels.append(msg)
model_results = results
model_names = names
return render_template('fit.html',
y_var_select=y_var_select,
pred_type_select=pred_type_select,
multiselect=multiselect,
model_results=allmodels,
model_names=names)