ML之catboost：catboost的CatBoostRegressor函数源代码简介、解读之详细攻略

冷不防 2022-11-19 01:13 389阅读 0赞

**目录**

catboost的CatBoostRegressor函数源代码简介、解读

--------------------

# catboost的CatBoostRegressor函数源代码简介、解读 #

<table> 
 <tbody> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> class CatBoostRegressor&nbsp;Found at: catboost.core class CatBoostRegressor(CatBoost): &nbsp;&nbsp;&nbsp;&nbsp;_estimator_type = 'regressor' &nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;Implementation of the scikit-learn API for CatBoost regression. &nbsp;&nbsp;&nbsp;&nbsp;Parameters &nbsp;&nbsp;&nbsp;&nbsp;---------- &nbsp;&nbsp;&nbsp;&nbsp;Like in CatBoostClassifier, except loss_function, classes_count, class_names and class_weights &nbsp;&nbsp;&nbsp;&nbsp;loss_function : string, [default='RMSE'] &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'RMSE' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'MAE' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'Quantile:alpha=value' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'LogLinQuantile:alpha=value' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'Poisson' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'MAPE' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'Lq:q=value' &nbsp;&nbsp;&nbsp;&nbsp;""" </td> 
 <td style="vertical-align:top;width:409px;"> 实现scikit-learn API的CatBoost回归。 参数 ---------- 像CatBoostClassifier，除了loss_function, classes_count, class_names和class_weights </td> 
 </tr> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> &nbsp;&nbsp;&nbsp;&nbsp;def __init__( &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;self, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;iterations=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;learning_rate=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;depth=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;l2_leaf_reg=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;model_size_reg=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;rsm=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;loss_function='RMSE', &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;border_count=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;feature_border_type=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;per_float_feature_quantization=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;input_borders=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;output_borders=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;fold_permutation_block=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;od_pval=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;od_wait=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;od_type=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;nan_mode=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;counter_calc_method=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;leaf_estimation_iterations=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;leaf_estimation_method=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;thread_count=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;random_seed=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;use_best_model=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;best_model_min_trees=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;silent=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;logging_level=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;metric_period=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ctr_leaf_count_limit=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;store_all_simple_ctr=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;max_ctr_complexity=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;has_time=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;allow_const_label=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;target_border=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;one_hot_max_size=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;random_strength=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;name=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ignored_features=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;train_dir=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;custom_metric=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;eval_metric=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;bagging_temperature=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;save_snapshot=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;snapshot_file=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;snapshot_interval=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;fold_len_multiplier=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;used_ram_limit=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gpu_ram_part=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;pinned_memory_size=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;allow_writing_files=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;final_ctr_computation_mode=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;approx_on_full_history=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;boosting_type=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;simple_ctr=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;combinations_ctr=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;per_feature_ctr=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ctr_description=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ctr_target_border_count=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;task_type=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;device_config=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;devices=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;bootstrap_type=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;subsample=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;mvs_reg=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;sampling_frequency=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;sampling_unit=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;dev_score_calc_obj_block_size=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;dev_efb_max_buckets=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;sparse_features_conflict_fraction=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;max_depth=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;n_estimators=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;num_boost_round=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;num_trees=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;colsample_bylevel=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;random_state=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;reg_lambda=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;objective=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;eta=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;max_bin=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gpu_cat_features_storage=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;data_partition=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;metadata=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;early_stopping_rounds=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;cat_features=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;grow_policy=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;min_data_in_leaf=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;min_child_samples=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;max_leaves=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;num_leaves=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;score_function=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;leaf_estimation_backtracking=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ctr_history_unit=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;monotone_constraints=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;feature_weights=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;penalties_coefficient=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;first_feature_use_penalties=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;per_object_feature_penalties=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;model_shrink_rate=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;model_shrink_mode=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;langevin=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;diffusion_temperature=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;posterior_sampling=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;boost_from_average=None): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;params = {} &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;not_params = ["not_params", "self", "params", "__class__"] &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;for key, value in iteritems(locals().copy()): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if key not in not_params and value is not None: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;params[key] = value &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;super(CatBoostRegressor, self).__init__(params) </td> 
 <td style="vertical-align:top;width:409px;"> </td> 
 </tr> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> &nbsp;&nbsp;&nbsp;&nbsp;def fit(self, X, y=None, cat_features=None, sample_weight=None, baseline=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;use_best_model=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;eval_set=None, verbose=None, logging_level=None, plot=False, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;column_description=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose_eval=None, metric_period=None, silent=None, early_stopping_rounds=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;save_snapshot=None, snapshot_file=None, snapshot_interval=None, init_model=None): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Fit the CatBoost model. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Parameters &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;---------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;X&nbsp;: catboost.Pool or list or numpy.ndarray or pandas.DataFrame or pandas.Series. If not catboost.Pool, 2 dimensional Feature matrix or string - file with dataset. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;y&nbsp;: list or numpy.ndarray or pandas.DataFrame or pandas.Series, optional (default=None). Labels, 1 dimensional array like.&nbsp;Use only if X is not catboost.Pool. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;cat_features&nbsp;: list or numpy.ndarray, optional (default=None). If not None, giving the list of Categ columns indices.Use only if X is not catboost.Pool. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;sample_weight : list or numpy.ndarray or pandas.DataFrame or pandas.Series, optional (default=None). Instance weights, 1 dimensional array like. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;baseline : list or numpy.ndarray, optional (default=None). If not None, giving 2 dimensional array like data.&nbsp;Use only if X is not catboost.Pool. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;use_best_model&nbsp;: bool, optional (default=None). Flag to use best model &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;eval_set&nbsp;: catboost.Pool or list, optional (default=None). A list of (X, y) tuple pairs to use as a validation set for&nbsp;early-stopping &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;metric_period&nbsp;: int. Frequency of evaluating metrics. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose&nbsp;: bool or int. If verbose is bool, then if set to True, logging_level is set to Verbose,&nbsp;if set to False, logging_level is set to Silent.&nbsp;If verbose is int, it determines the frequency of writing metrics to output and&nbsp;logging_level is set to Verbose. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;silent : bool. If silent is True, logging_level is set to Silent.&nbsp;&nbsp;If silent is False, logging_level is set to Verbose. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;logging_level&nbsp;: string, optional (default=None). Possible values: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;- 'Silent' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;- 'Verbose' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;- 'Info' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;- 'Debug' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;plot : bool, optional (default=False). If True, draw train and eval error in Jupyter notebook &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose_eval&nbsp;: bool or int. &nbsp;Synonym for verbose. Only one of these parameters should be set. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;early_stopping_rounds&nbsp;: int. Activates Iter overfitting detector with od_wait set to&nbsp;early_stopping_rounds. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;save_snapshot : bool, [default=None]. Enable progress snapshotting for restoring progress after crashes or interruptions &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;snapshot_file : string, [default=None]. Learn progress snapshot file path, if None will use default filename&nbsp;&nbsp;snapshot_interval: int, [default=600]. Interval between saving snapshots (seconds) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;init_model&nbsp;: CatBoost class or string, [default=None]. Continue training starting from the existing model.&nbsp;If this parameter is a string, load initial model from the path specified by this string. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Returns &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;model : CatBoost &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;params = deepcopy(self._init_params) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;_process_synonyms(params) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if 'loss_function' in params: &nbsp;&nbsp; </td> 
 <td style="vertical-align:top;width:409px;"> X: catboost。pool或list或numpy。ndarray或pandas.DataFrame或pandas.Series。如果不是catboost。Pool，二维特征矩阵或字符串文件与数据集。 y: list或numpy。ndarray或pandas.DataFrame或pandas.Series。可选(默认= None)。标签，类似于一维数组。仅当X不是catboost.Pool时使用。 cat_features:&nbsp;list或numpy.ndarray,可选(默认= None)。如果不是None，则给出类别列索引的列表。仅当X不是catboost.Pool时使用。 sample_weight:列表或numpy。ndarray或pandas.DataFrame或pandas.Series,可选(默认= None)。实例权重，类似于一维数组。 baseline:列表或numpy。ndarray,可选(默认= None)。如果不是None，则给出像data这样的二维数组。仅当X不是catboost.Pool时使用。 use_best_model: bool，可选(默认为None)。标记使用最佳模型 eval_set: catboost。Pool或列表，可选(默认为None)。(X, y)元组对的列表，用作早期停止的验证集。 metric_period: int。评估指标的频率。 verbose: bool或int。如果verbose是bool，那么如果设置为True, logging_level将设置为verbose，如果设置为False, logging_level将设置为Silent。如果verbose为int，则它确定向输出写入指标的频率，并将logging_level设置为verbose。 silent :&nbsp;bool。如果silent为True, loging_level设置为silent。如果silent为False, loging_level设置为Verbose。 logging_level:字符串，可选(默认为None)。可能的值: ——“沉默” ——“详细” ——“信息” ——“调试” plot: bool，可选(默认=False)。如果为真，在Jupyter中绘制训练集和测试集的error verbose_eval: bool或int。详细的同义词。应该只设置这些参数中的一个。 early_stopping_rounds: int。激活Iter过拟合检测器，od_wait设置为early_stopping_rounds。 save_snapshot:&nbsp;bool， [default=None]。启用进度快照，以便在崩溃或中断后恢复进度 snapshot_file: string， [default=None]。学习进度快照文件路径，如果没有将使用默认文件名snapshot_interval: int，[默认=600]。保存快照的时间间隔(秒) init_model: CatBoost类或字符串，[default=None]。从现有的模式开始继续培训。如果该参数为字符串，则从该字符串指定的路径加载初始模型。 </td> 
 </tr> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> self._check_is_regressor_loss(params['loss_function']) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return self._fit(X, y, cat_features, None, None, None, sample_weight, None, None, None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;None, baseline, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;use_best_model, eval_set, verbose, logging_level, plot, column_description, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose_eval, metric_period, silent, early_stopping_rounds, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;save_snapshot, snapshot_file, snapshot_interval, init_model) &nbsp;&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp;&nbsp;def predict(self, data, prediction_type=None, ntree_start=0, ntree_end=0, thread_count=- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;1, verbose=None): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Predict with data. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Parameters &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;---------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;data&nbsp;: catboost.Pool or list of features or list of lists or numpy.ndarray or pandas.&nbsp;DataFrame or pandas.Series&nbsp;or catboost.FeaturesData.&nbsp;Data to apply model on.&nbsp;If data is a simple list (not list of lists) or a one-dimensional numpy.ndarray it is interpreted&nbsp;as a list of features for a single object. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;prediction_type&nbsp;: string, optional (default='RawFormulaVal'). &nbsp;Can be: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;- 'RawFormulaVal' : return raw formula value. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;- 'Exponent' : return Exponent of raw formula value. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_start: int, optional (default=0) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Model is applied on the interval [ntree_start, ntree_end) (zero-based indexing). &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_end: int, optional (default=0) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Model is applied on the interval [ntree_start, ntree_end) (zero-based indexing).&nbsp;If value equals to 0 this parameter is ignored and ntree_end equal to tree_count_. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;thread_count : int (default=-1). The number of threads to use when applying the model.&nbsp;Allows you to optimize the speed of execution. This parameter doesn't affect results.&nbsp;If -1, then the number of threads is set to the number of CPU cores. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose : bool. If True, writes the evaluation metric measured set to stderr. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Returns &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;prediction :&nbsp;If data is for a single object, the return value is single float formula return value&nbsp;otherwise one-dimensional numpy.ndarray of formula return values for each object. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if prediction_type is None: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;prediction_type = self._get_default_prediction_type() &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return self._predict(data, prediction_type, ntree_start, ntree_end, thread_count, verbose, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;'predict') </td> 
 <td style="vertical-align:top;width:409px;"> 参数 --------- data&nbsp;: catboost。池或特性列表或列表的列表或numpy。ndarray或熊猫。DataFrame或熊猫。系列或catboost.FeaturesData。应用模型的数据。如果data是一个简单的列表(不是列表的列表)或一维numpy。ndarray它被解释为一个对象的特性列表。 prediction_type&nbsp;:字符串，可选(默认为'RawFormulaVal')。可以是: - 'RawFormulaVal':返回原始公式值。 - 'Exponent':返回原始公式值的指数。 ntree_start: int，可选(默认为0) 模型应用于区间[ntree_start, ntree_end)(从零开始索引)。 ntree_end:&nbsp; int，可选(默认为0) 模型应用于区间[ntree_start, ntree_end)(从零开始索引)。如果value等于0，则忽略该参数，ntree_end等于tree_count_。 thread_count :int(默认=-1)。应用模型时要使用的线程数。允许您优化执行速度。此参数不影响结果。如果-1，则线程数设置为CPU核数。 verbose :bool。如果为真，则将评估度量值写入stderr。 返回 ------- prediction：如果数据是针对单个对象的，则返回值为单个float公式返回值，否则为一维numpy。ndarray的公式返回每个对象的值。 </td> 
 </tr> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> &nbsp;&nbsp;&nbsp;&nbsp;def staged_predict(self, data, prediction_type='RawFormulaVal', ntree_start=0, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_end=0, eval_period=1, thread_count=-1, verbose=None): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Predict target at each stage for data. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Parameters &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;---------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;data : catboost.Pool or list of features or list of lists or numpy.ndarray or pandas.&nbsp;DataFrame or pandas.Series&nbsp;or catboost.FeaturesData. Data to apply model on.&nbsp;If data is a simple list (not list of lists) or a one-dimensional numpy.ndarray it is interpreted&nbsp;as a list of features for a single object. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_start: int, optional (default=0). Model is applied on the interval [ntree_start, ntree_end) with the step eval_period (zero-based indexing). &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_end: int, optional (default=0).Model is applied on the interval [ntree_start, ntree_end) with the step eval_period (zero-based indexing).&nbsp;If value equals to 0 this parameter is ignored and ntree_end equal to tree_count_. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;eval_period: int, optional (default=1). Model is applied on the interval [ntree_start, ntree_end) with the step eval_period (zero-based indexing). &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;thread_count : int (default=-1). The number of threads to use when applying the model.&nbsp;Allows you to optimize the speed of execution. This parameter doesn't affect results.&nbsp;If -1, then the number of threads is set to the number of CPU cores. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose&nbsp;: bool. If True, writes the evaluation metric measured set to stderr. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Returns &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;prediction : generator for each iteration that generates:If data is for a single object, the return value is single float formula return value&nbsp;&nbsp;otherwise one-dimensional numpy.ndarray of formula return values for each object. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return self._staged_predict(data, prediction_type, ntree_start, ntree_end, eval_period, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;thread_count, verbose, 'staged_predict') </td> 
 <td style="vertical-align:top;width:409px;"> &nbsp;&nbsp;data : catboost。池或特性列表或列表的列表或numpy。ndarray或DataFrame 或pandas.Series&nbsp;or catboost.FeaturesData。应用模型的数据。如果data是一个简单的列表(不是列表的列表)或一维numpy。ndarray它被解释为一个对象的特性列表。 ntree_start: int，可选(默认为0)。模型应用于间隔[ntree_start, ntree_end)，步长为eval_period(从零开始索引)。 ntree_end:int，可选(默认为0)。模型应用于间隔[ntree_start, ntree_end)，步长为eval_period(从零开始索引)。如果value等于0，则忽略该参数，ntree_end等于tree_count_。 eval_period:&nbsp;&nbsp;int，可选(默认为1)。模型应用于间隔[ntree_start, ntree_end)，步长为eval_period(从零开始索引)。 thread_count : int(默认=-1)。应用模型时要使用的线程数。允许您优化执行速度。此参数不影响结果。如果-1，则线程数设置为CPU核数。 verbose&nbsp;:bool。如果为真，则将评估度量值写入stderr。 返回 ------- prediction :为每个迭代生成的生成器:如果数据是针对单个对象的，则返回值为单个float公式返回值，否则为一维numpy。ndarray的公式返回每个对象的值。 </td> 
 </tr> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> &nbsp;&nbsp;&nbsp;&nbsp;def score(self, X, y=None): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Calculate R^2. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Parameters &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;---------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;X&nbsp;: catboost.Pool or list or numpy.ndarray or pandas.DataFrame or pandas.Series.Data to apply model on. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;y : list or numpy.ndarray.True labels. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Returns &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;------- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;R^2 : float &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;""" &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if isinstance(X, Pool): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if y is not None: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;raise CatBoostError("Wrong initializing y: X is catboost.Pool object, y must be &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;initialized inside catboost.Pool.") &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;y = X.get_label() &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if y is None: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;raise CatBoostError("Label in X has not initialized.") &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;elif y is None: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;raise CatBoostError("y should be specified.") &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;y = np.array(y, dtype=np.float64) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;predictions = self._predict(X, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;prediction_type=self._get_default_prediction_type(), &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_start=0, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;ntree_end=0, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;thread_count=-1, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;verbose=None, &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;parent_method_name='score') &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;loss = self._object._get_loss_function_name() &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if loss == 'RMSEWithUncertainty': &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;predictions = predictions[:0] &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;total_sum_of_squares = np.sum((y - y.mean(axis=0)) ** 2) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;residual_sum_of_squares = np.sum((y - predictions) ** 2) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return 1 - residual_sum_of_squares / total_sum_of_squares </td> 
 <td style="vertical-align:top;width:409px;"> </td> 
 </tr> 
 <tr> 
 <td style="vertical-align:top;width:495px;"> &nbsp;&nbsp;&nbsp;&nbsp;def _check_is_regressor_loss(self, loss_function): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;is_regression = self._is_regression_objective(loss_function) or self. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;_is_multiregression_objective(loss_function) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if isinstance(loss_function, str) and not is_regression: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;raise CatBoostError("Invalid loss_function='{}': for regressor use " &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;"RMSE, MultiRMSE, MAE, Quantile, LogLinQuantile, Poisson, MAPE, Lq or custom &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;objective object".format(loss_function)) &nbsp;&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp;&nbsp;def _get_default_prediction_type(self): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;# TODO(ilyzhin) change on get_all_params after MLTOOLS-4758 &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;params = deepcopy(self._init_params) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;_process_synonyms(params) &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;loss_function = params.get('loss_function') &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if loss_function and isinstance(loss_function, str): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if loss_function.startswith('Poisson') or loss_function.startswith('Tweedie'): &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return 'Exponent' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if loss_function == 'RMSEWithUncertainty': &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return 'RMSEWithUncertainty' &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return 'RawFormulaVal' </td> 
 <td style="vertical-align:top;width:409px;"> </td> 
 </tr> 
 </tbody> 
</table>