ScitypeHierarchical

class ScitypeHierarchical(is_univariate=None, is_equally_spaced=None, is_equal_length=None, is_empty=None, is_one_series=None, is_one_panel=None, has_nans=None, n_instances=None, n_panels=None, n_features=None, feature_names=None, dtypekind_dfip=None, feature_kind=None)[source]

Hierarchical data type. Represents a hierarchical collection of time series.

The Hierarchical data type is an abstract data type (= scitype).

A Hierarchical represents a hierarchically indexed collection of many single monotonously indexed sequences, that is, an indexed collection of objects that follow the Series data type, where the index is hierarchical (multi-level).

This including the sub-case of what is commonly called a “hierarchical time series”, if the index is interpreted as time.

Formally, an abstract Hierarchical object that contains \(N\) time series has:

an index \(s_1, \dots, s_N\), where each \(s_i\) is an \(H\)-tuple of integer or a categorical types, representing the series identifier, also called “hierarchical index”
individual time series \(S_i\) for \(i = 1, \dots, N\), where each \(S_i\) is an object of abstract Series type
above, the domain of the hierarchical index is a fixed Cartesian product of domains, \(\mathcal{S} = \mathcal{S}_1 \times \dots \times \mathcal{S}_H\) where each \(\mathcal{S}_h\) is an integer or categorical type domain.

To be considered of Hierarchical type, there must be at least two hierarchy levels, i.e., \(H \geq 2\). If \(H = 1\), the data is considered of Panel type (to avoid duplicate typing).

The object \(S_i\) is interpreted as the time series (or sequence) at the instance index \(s_i\).

The indices \(s_i\) are assumed distinct, but not necessarily ordered.

The domains \(\mathcal{S}_h\) are interpreted as hierarchy levels.

Concrete types implementing the Hierarchical data type must specify:

hierarchy: how the hierarchy levels are represented
hierarchy names: optional, names of the hierarchy levels
variables: how the dimensions of the individual time series are represented
variable names: optional, names of the variable dimensions
time points: how time points are represented in the individual time series
time index: how the time index is represented

Concrete implementations may implement only sub-cases of the full abstract type.

Parameters:

is_univariate: bool: True iff table has one variable
is_equally_spacedbool: True iff all series in the hierarchical collection are equally spaced
is_equal_length: bool: True iff all series in the hierarchical collection are of equal length
is_empty: bool: True iff table has no variables or no instances
is_one_series: bool: True iff there is only one series in the hierarchical collection.
is_one_panel: bool: True iff there is only one flat panel in the hierarchical collection, i.e., the collection has a flat hierarchy, plus additional hierarchy levels that have only a single value.
has_nans: bool: True iff the hierarchical series contains NaN values
n_instances: int: number of instances, i.e., individual time series, in the hierarchical collection. Formally, the number of unique hierarchy indices \(s_i\), namely \(N\), in the definition above.
n_panels: int: number of flat panels in the hierarchical collection. If the indices are \(H\)-tuples, this is the number of unique values of the \((H-1)\)-tuples obtained by removing the last entry of each \(s_i\).
n_features: int: number of variables in the hierarchical series
feature_names: list of int or object: names of variables in the hierarchical series
dtypekind_dfip: list of DtypeKind enum: list of DtypeKind enum values for each feature in the hierarchical series, following the data frame interface protocol. In same order as feature_names.
feature_kind: list of str: list of feature-kind strings for each feature in the hierarchical series, coerced to "FLOAT" or "CATEGORICAL" type string. In same order as feature_names.

Methods

`__call__`(obj[, return_metadata, var_name, ...])	Check if obj is of this data type.
`check`(obj[, return_metadata, var_name, ...])	Check if obj is of this data type.
`clone`()	Obtain a clone of the object with same hyper-parameters and config.
`clone_tags`(estimator[, tag_names])	Clone tags from another object as dynamic override.
`create_test_instance`([parameter_set])	Construct an instance of the class, using first test parameter set.
`create_test_instances_and_names`([parameter_set])	Create list of all test instances and a list of names for them.
`get`(key[, default])	Get attribute by key.
`get_class_tag`(tag_name[, tag_value_default])	Get class tag value from class, with tag level inheritance from parents.
`get_class_tags`()	Get class tags from class, with tag level inheritance from parent classes.
`get_config`()	Get config flags for self.
`get_param_defaults`()	Get object's parameter defaults.
`get_param_names`([sort])	Get object's parameter names.
`get_params`([deep])	Get a dict of parameters values for this object.
`get_tag`(tag_name[, tag_value_default, ...])	Get tag value from instance, with tag level inheritance and overrides.
`get_tags`()	Get tags from instance, with tag level inheritance and overrides.
`get_test_params`([parameter_set])	Return testing parameter settings for the skbase object.
`is_composite`()	Check if the object is composed of other BaseObjects.
`load_from_path`(serial)	Load object from file location.
`load_from_serial`(serial)	Load object from serialized memory container.
`reset`()	Reset the object to a clean post-init state.
`save`([path, serialization_format])	Save serialized self to bytes-like object or to (.zip) file.
`set_config`(**config_dict)	Set config flags to given values.
`set_params`(**params)	Set the parameters of this object.
`set_random_state`([random_state, deep, ...])	Set random_state pseudo-random seed parameters for self.
`set_tags`(**tag_dict)	Set instance level tag overrides to given values.