I determined the settings for a Seasonal AutoRegressive Integrated Moving Average (SARIMA) model to forecast the “hotel_avg_daily_rate” time series. This process involved acknowledging the data’s clear seasonality and using standard methods like analyzing autocorrelation through ACF and PACF plots. The choice of parameters, including ARIMA (p, d, q) and seasonal elements (P, D, Q, S), was guided by these plots.
After selecting the parameters, I implemented the SARIMA model on the training data. I then validated the model by making predictions on the test dataset and comparing these predictions with the actual figures. I produced forecasts for the upcoming 12 months, including projected values and 95% confidence intervals, and calculated the Root Mean Squared Error (RMSE) to assess the model’s prediction accuracy. With an RMSE of around 13.12, this metric helps gauge the model’s fit to the data, where a lower RMSE indicates a more accurate model.
Be First to Comment