Prepare data to fit the TFT model. Args: df: Original DataFrame. fillna: Whether to fill the data with the mean values. Returns: Transformed DataFrame.
(df, dataset, fillna=False)
| 92 | |
| 93 | |
| 94 | def process_qlib_data(df, dataset, fillna=False): |
| 95 | """Prepare data to fit the TFT model. |
| 96 | |
| 97 | Args: |
| 98 | df: Original DataFrame. |
| 99 | fillna: Whether to fill the data with the mean values. |
| 100 | |
| 101 | Returns: |
| 102 | Transformed DataFrame. |
| 103 | |
| 104 | """ |
| 105 | # Several features selected manually |
| 106 | feature_col = DATASET_SETTING[dataset]["feature_col"] |
| 107 | label_col = [DATASET_SETTING[dataset]["label_col"]] |
| 108 | temp_df = df.loc[:, feature_col + label_col] |
| 109 | if fillna: |
| 110 | temp_df = fill_test_na(temp_df) |
| 111 | temp_df = temp_df.swaplevel() |
| 112 | temp_df = temp_df.sort_index() |
| 113 | temp_df = temp_df.reset_index(level=0) |
| 114 | dates = pd.to_datetime(temp_df.index) |
| 115 | temp_df["date"] = dates |
| 116 | temp_df["day_of_week"] = dates.dayofweek |
| 117 | temp_df["month"] = dates.month |
| 118 | temp_df["year"] = dates.year |
| 119 | temp_df["const"] = 1.0 |
| 120 | return temp_df |
| 121 | |
| 122 | |
| 123 | def process_predicted(df, col_name): |
no test coverage detected