A student uses the following code to fit and evaluate Holt-Winters forecasting on monthly demand data:

from statsmodels.tsa.holtwinters import ExponentialSmoothing from sklearn.metrics import mean_squared_error model = ExponentialSmoothing( data, trend="add", seasonal="add", seasonal_periods=12 ).fit() forecast = model.forecast(6) rmse = mean_squared_error(data[-6:], forecast) ** 0.5
Assume data contains 60 monthly observations, and the objective is to evaluate the model on the final six months. What is the most serious methodological problem?

Question

A student uses the following code to fit and evaluate Holt-Winters forecasting on monthly demand data: 
 
 from statsmodels.tsa.holtwinters import ExponentialSmoothing from sklearn.metrics import mean_squared_error model = ExponentialSmoothing(     data,     trend="add",     seasonal="add",     seasonal_periods=12 ).fit() forecast = model.forecast(6) rmse = mean_squared_error(data[-6:], forecast) ** 0.5 
  Assume  data  contains 60 monthly observations, and the objective is to evaluate the model on the final six months. What is the most serious methodological problem?

Accepted Answer

The model was fitted using the observations that are subsequently treated as test data.

Answer

The seasonal period should be 6 because six observations are being forecast.

Answer

RMSE cannot be computed for Holt-Winters forecasts.

Answer

The model should have used a multiplicative seasonal component.

Crowdly

A student uses the following code to fit and evaluate Holt-Winters forecasting o...

Want instant access to all verified answers on moodle.spit.ac.in?