Evaluate for training data. Parameters ---------- feval : callable or None, optional (default=None) Customized evaluation function. Should accept two parameters: preds, train_data, and return (eval_name, eval_result, is_higher_better) or l
(self, feval=None)
| 2163 | return self.__inner_eval(name, data_idx, feval) |
| 2164 | |
| 2165 | def eval_train(self, feval=None): |
| 2166 | """Evaluate for training data. |
| 2167 | |
| 2168 | Parameters |
| 2169 | ---------- |
| 2170 | feval : callable or None, optional (default=None) |
| 2171 | Customized evaluation function. |
| 2172 | Should accept two parameters: preds, train_data, |
| 2173 | and return (eval_name, eval_result, is_higher_better) or list of such tuples. |
| 2174 | |
| 2175 | preds : list or numpy 1-D array |
| 2176 | The predicted values. |
| 2177 | train_data : Dataset |
| 2178 | The training dataset. |
| 2179 | eval_name : string |
| 2180 | The name of evaluation function (without whitespaces). |
| 2181 | eval_result : float |
| 2182 | The eval result. |
| 2183 | is_higher_better : bool |
| 2184 | Is eval result higher better, e.g. AUC is ``is_higher_better``. |
| 2185 | |
| 2186 | For multi-class task, the preds is group by class_id first, then group by row_id. |
| 2187 | If you want to get i-th row preds in j-th class, the access way is preds[j * num_data + i]. |
| 2188 | |
| 2189 | Returns |
| 2190 | ------- |
| 2191 | result : list |
| 2192 | List with evaluation results. |
| 2193 | """ |
| 2194 | return self.__inner_eval(self._train_data_name, 0, feval) |
| 2195 | |
| 2196 | def eval_valid(self, feval=None): |
| 2197 | """Evaluate for validation data. |