Python for Finance Cookbook

There's more...

We are also interested in some basic performance evaluation of the fitted model. Execute the following steps to visually inspect the predicted versus actual gold prices in 2005.

1. Merge the test set with the forecasts:

selected_columns = ['ds', 'yhat_lower', 'yhat_upper', 'yhat']df_pred = df_pred.loc[:, selected_columns].reset_index(drop=True)df_test = df_test.merge(df_pred, on=['ds'], how='left')df_test.ds = pd.to_datetime(df_test.ds)df_test.set_index('ds', inplace=True)

We merged the test set with the prediction DataFrame. We used a left join, which returns all the rows from the left table (test set) and the matched rows from the right table (prediction DataFrame) while leaving the unmatched rows empty. This way, we also ...