bill2watt.predictors package
Submodules
bill2watt.predictors.base_predictor module
This module contains the ‘BasePredictor’ class, which is used as a base for the models that predict typical load profiles based on monthly time-of-use (ToU) energy consumption values.
Notes
Info
Author: G. Lorenti Email: gianmarco.lorenti@polito.it
- class bill2watt.predictors.base_predictor.BasePredictor(x_data=None, y_data=None)
Bases:
objectBase class for predictors of typical load profiles (y) based on monthly time-of-use (ToU) energy consumption values (x).
Parameters
- x_datapandas.DataFrame or None, optional
ToU energy consumption values associated with the SLPs.
- y_datapandas.DataFrame or None, optional
Typical load profiles of the Standard Load Profiles (SLPs).
Attributes
- x_data (property)pd.DataFrame or None
ToU energy consumption values of the data points used to build the predictor. None for predictors that do not require data points.
- y_data (property)pd.DataFrame or None
Typical load profiles of the data points used to build the predictor. None for predictors that do not require data points.
- n (property)int
Number of data points used to build the predictor.
Methods
- add_data(x_data, y_data, update_existing=True)
Add new data points to the predictor.
- predict(x, nd)
Placeholder method that must be implemented by child classes.
- property x_data
Getter for the x_data property.
Returns a copy of the x_data if it is not None, otherwise returns None.
Returns
- pd.DataFrame or None
Copy of the x_data or None.
- property y_data
Getter for the y_data property.
Returns a copy of the y_data if it is not None, otherwise returns None.
Returns
- pd.DataFrame or None
Copy of the y_data or None.
- property n
Getter for the n property.
Returns the number of data points (length of x_data or y_data). Returns None if both x_data and y_data are None.
Returns
- int
Number of data points.
- add_data(x_data=None, y_data=None, update_existing=True)
Add new data points to the predictor.
Parameters
- x_datapd.DataFrame or None
ToU energy consumption values of the new data points. None if no data points are provided.
- y_datapd.DataFrame or None
Typical load profiles of the new data points. None if no data points are provided.
- update_existingbool, optional
Flag indicating whether to update existing data points (i.e., already in the index). Default is True.
Returns
- pd.DataFrame or None
The updated x_data after adding new data points.
- pd.DataFrame or None
The updated y_data after adding new data points.
Raises
- AssertionError
If ‘x_data’ and ‘y_data’ do not comply with requirements for new data points.
- predict(x, nd)
bill2watt.predictors.category_predictor module
This module contains the ‘CategoryPredictor’ class, which is used for predicting typical load profiles based on monthly time-of-use (ToU) energy consumption values, using training data and an a-priori categorization approach.
Notes
Info
Author: G. Lorenti Email: gianmarco.lorenti@polito.it
- class bill2watt.predictors.category_predictor.CategoryPredictor(x_data=None, y_data=None, category=None)
Bases:
BasePredictorThis class implements a predictor for evaluating typical load profiles based on monthly time-of-use (ToU) energy consumption values, using categories defined by a specific key.
Parameters
- x_datapandas.DataFrame or None
ToU energy consumption values associated with the training points. If None, default values are used.
- y_datapandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
- categorystr or None, optional
The name of the column/index level used for categorization. If None and ‘y_data’ is not None, the name of the index of ‘y_data’ is used.
Attributes
- x_data (property)pandas.DataFrame or None
ToU energy consumption values of the training points. If None, default values are used.
- y_data (property)pandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
- category (property)str or None
The name of the column/index level used for categorization.
- y_categories (property)pandas.DataFrame or None
Categorized typical load profiles based on the training data. If None, no categorization has been performed yet.
- categories (property)list or None
Unique category values present in the categorized data. If None, no categorization has been performed yet.
Methods
- predict(x, key)
Evaluate typical load profiles based on ToU energy consumption values and using the categories.
- add_data(x_data, y_data, **kwargs)
Add new data points to the predictor and re-train the categorizer.
See Also
BasePredictor : Base class for typical load profile predictors.
- property category
Get or set the name of the column/index level used for categorization.
Returns
- str or None
The name of the column/index level used for categorization. If None, no categorization has been performed yet.
- property y_categories
Categorized typical load profiles based on the training data.
Returns
- pandas.DataFrame or None
Categorized typical load profiles. If None, no categorization has been performed yet.
- property categories
Unique category values present in the categorized data.
Returns
- list or None
Unique category values. If None, no categorization has been performed yet.
- predict(x, key)
Evaluate typical load profiles based on ToU energy consumption values and using the categories.
Parameters
- xarray-like
ToU energy consumption values of the point(s) to predict.
- keytuple or list of tuples
Key values related to the categorized data associated with each point.
Returns
- np.ndarray
Predicted typical load profiles based on the provided inputs.
Raises
- AssertionError
If any inconsistency is found in the input data or key values are not present in the categorized data.
- add_data(x_data, y_data, **kwargs)
Add new training points to the predictor and re-train the categorizer.
Parameters
- x_dataarray-like
ToU energy consumption values of the new data points.
- y_dataarray-like
Typical load profiles of the new data points.
Additional Parameters
Parameters from the ‘add_data’ method of the ‘BasePredictor’ parent class.
Raises
- AssertionError
If ‘x_data’ and ‘y_data’ do not comply with the requirements for new data points.
bill2watt.predictors.classification_predictor module
This module contains the ‘ClassificationPredictor’ class, which is used for predicting typical load profiles based on monthly time-of-use (ToU) energy consumption values, using training data and a classification approach.
Notes
Info
Author: G. Lorenti Email: gianmarco.lorenti@polito.it
- class bill2watt.predictors.classification_predictor.ClassificationPredictor(x_data=None, y_data=None, clustering=None, classifier=None)
Bases:
BasePredictorThis class implements a predictor for evaluating typical load profiles based on time-of-use (ToU) energy consumption values using training data in a classification and clustering approach.
Parameters
- x_datapandas.DataFrame or None
ToU energy consumption values of the training points. If None, default values are used.
- y_datapandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
- clusteringobject
Clustering model object with fit and predict methods.
- classifierobject
Classification model object with fit and predict methods.
Attributes
- x_data (property)pandas.DataFrame or None
ToU energy consumption values of the training points. If None, default values are used.
- y_data (property)pandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
Methods
- predict(x)
Predict typical load profiles based on the provided inputs.
- add_data(x, y)
Add new data points to the predictor.
See Also
BasePredictor : Base class for typical load profile predictors.
- predict(x)
Predict typical load profiles based on the provided inputs.
Parameters
- xarray-like
ToU energy consumption values of the point(s) to predict.
Returns
- np.ndarray
Predicted typical load profiles based on the provided inputs.
Notes
If ‘x’ refers to a single point, then the output is provided as a single point. If ‘x’ refers to multiple points, the output is provided as a matrix with the same number of rows as ‘x’.
Raises
- AssertionError
If any inconsistency is found in the input data.
- add_data(x_data, y_data, **kwargs)
Add new training points to the predictor and re-train the models.
Parameters
- x_dataarray-like
ToU energy consumption values of the new data points.
- y_dataarray-like
Typical load profiles of the new data points.
Additional parameters
Parameters from ‘add_data’ method of ‘BasePredictor’ parent class.
Raises
- AssertionError
If ‘x_data’ and ‘y_data’ do not comply with requirements for new data points.
bill2watt.predictors.flat_predictor module
Flat Predictor
Description
This module contains the FlatPredictor class, which is used for predicting typical load profiles based on monthly time-of-use (ToU) energy consumption values, using a flat approach that spreads the consumption in each ToU tariff.
Notes
None.
Info
Author: G. Lorenti Email: gianmarco.lorenti@polito.it
- class bill2watt.predictors.flat_predictor.FlatPredictor
Bases:
BasePredictorFlat Predictor.
This class implements a predictor for evaluating typical load profiles based on monthly time-of-use (ToU) energy consumption values using a flat approach that spreads the consumption in each ToU tariff.
Parameters
Attributes
Methods
- predict(x, nd)
Evaluate typical load profiles based on energy consumption values and number of days of each day-type.
See Also
BasePredictor : Base class for typical load profile predictors.
- predict(x, nd)
Evaluate typical load profiles in multiple points, based on ToU energy consumption values, using a flat approach.
Parameters
- xarray-like
ToU energy consumption values of the point(s) to predict.
- ndarray-like
Number of days of each day-type for each point.
Returns
- np.ndarray
Predicted typical load profiles based on the provided inputs.
Notes
If ‘x’ (and other input parameters) refer to a single point, then the output is provided with a shape of a single point.
Raises
- AssertionError
If any inconsistency is found in the input data.
- add_data(x_data=None, y_data=None, **kwargs)
Add a data point to the predictor. This method is not supported by FlatPredictor and will raise an error unless ‘x_data’ and ‘y_data’ are both None.
Parameters
- xNone
Energy consumption values of the data point.
- yNone
Corresponding load profile of the data point.
- *+kwargsdict
Additional parameters, added for compatibility.
Raises
- NotImplementedError
This method is not supported by FlatPredictor and cannot be used unless ‘x_data’ and ‘y_data’ are both None.
bill2watt.predictors.regression_predictor module
This module contains the ‘RegressionPredictor’ class, which is used for predicting typical load profiles based on monthly time-of-use (ToU) energy consumption values, using training data and a regression approach.
Notes
Info
Author: G. Lorenti Email: gianmarco.lorenti@polito.it
- class bill2watt.predictors.regression_predictor.RegressionPredictor(x_data=None, y_data=None, regressor=None)
Bases:
BasePredictorThis class implements a predictor for evaluating typical load profiles based on time-of-use (ToU) energy consumption values using training data in a regression approach.
Parameters
- x_datapandas.DataFrame or None
ToU energy consumption values of the training points. Since they are not used, can be also None.
- y_datapandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
- regressorobject
Regression model object with fit and predict methods.
Attributes
- x_data (property)pandas.DataFrame or None
ToU energy consumption values of the training points. If None, default values are used.
- y_data (property)pandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
Methods
- predict(x)
Predict typical load profiles based on the provided inputs.
- add_data(x, y)
Add new data points to the predictor.
See Also
BasePredictor : Base class for typical load profile predictors.
- predict(x)
Predict typical load profiles based on the provided inputs.
Parameters
- xarray-like
ToU energy consumption values of the point(s) to predict.
Returns
- np.ndarray
Predicted typical load profiles based on the provided inputs.
Notes
If ‘x’ (and other input parameters) refer to a single point, then the output is provided with a shape of a single point.
Raises
- AssertionError
If any inconsistency is found in the input data.
- add_data(x_data, y_data, **kwargs)
Add new training points to the predictor and re-train the regressor.
Parameters
- x_dataarray-like
ToU energy consumption values of the new data points.
- y_dataarray-like
Typical load profiles of the new data points.
Additional parameters
Parameters from ‘add_data’ method of ‘BasePredictor’ parent class.
Raises
- AssertionError
If ‘x_data’ and ‘y_data’ do not comply with requirements for new data points.
bill2watt.predictors.slp_predictor module
This module contains the ‘SLPPredictor’ class, which is used for predicting typical load profiles based on monthly time-of-use (ToU) energy consumption values, using Standard Load Profiles (SPLs).
Notes
Info
Author: G. Lorenti Email: gianmarco.lorenti@polito.it
- class bill2watt.predictors.slp_predictor.SLPPredictor(x_data=None, y_data=None)
Bases:
BasePredictorThis class implements a predictor for evaluating typical load profiles based on monthly time-of-use (ToU) energy consumption values, using Standard Load Profiles (SLPs).
Parameters
- x_datapandas.DataFrame or None
ToU energy consumption values associated with the SLPs. Since they are not used, can be also None.
- y_datapandas.DataFrame or None
Typical load profiles of the Standard Load Profiles (SLPs). If None, default values are used.
Attributes
- x_data (property)pandas.DataFrame or None
ToU energy consumption values of the training points. If None, default values are used.
- y_data (property)pandas.DataFrame or None
Typical load profiles of the training points. If None, default values are used.
Methods
- predict(x, key, nd=None, scaler=None)
Evaluate typical load profiles based on ToU energy consumption values and using the SLPs.
- add_data(x, y)
Add new data points to the predictor.
See Also
BasePredictor : Base class for typical load profile predictors.
- predict(x, key)
Evaluate typical load profiles in multiple points, based on ToU energy consumption values, using the SLPs.
Parameters
- xarray-like
ToU energy consumption values of the point(s) to predict.
- keytuple or list of tuples
Key values related to the SLP associated with each point.
Returns
- np.ndarray
Predicted typical load profiles based on the provided inputs.
Notes
If ‘x’ (and other input parameters) refer to a single point, then the output is provided with a shape of a single point.
Raises
- AssertionError
If any inconsistency is found in the input data or key values are not present in the model data.