foehnix.Foehnix¶

class foehnix.Foehnix(predictor, data, concomitant=None, switch=False, filter_method=None, family='gaussian', control=None, **kwargs)[source]¶

Foehn Classification Based on a Two-Component Mixture Model

This is the main method of the foehnix package to estimate two-component mixture models for automated foehn classification.

__init__(predictor, data, concomitant=None, switch=False, filter_method=None, family='gaussian', control=None, **kwargs)[source]¶

Initialize parmeters which all methods need.

Parameters:

predictor : str Name of the main predictor (covariate) variable which is used to identify the foehn/no-foehn cluster. Must be present in data. data : pandas.DataFrame Index must be a time object, rows must contain neccesary data concomitant : str or list of str Name(s) of the covariates for the concomitant model. Must be present in data. If None (default), a mixture model without concomitants will be initialized. switch : bool False (default) if higher values of covariate y are assumed to be the foehn cluster. True if lower values are the foehn cluster. filter_method : dict, function or None Evaluates a filter on the data. E.g. a filter on the wind direction data to only use data from within a certain wind sector. See foehnix.foehnix_filter for details on the syntax. family : str or foehnix.Family class ‘gaussian’ (default) ‘logistic’ control : foehnix.foehnix.Control If None (default) it will be initialized. kwargs : kwargs to pass to the control function

Methods

__init__(predictor, data[, concomitant, …])	Initialize parmeters which all methods need.
no_concomitant_fit(y, control)	Fitting foehnix Mixture Model Without Concomitant Model.
plot(which, **kwargs)	Plotting method, helper function.
predict([newdata, returntype])	Predict method for foehnix Mixture Models
summary([detailed])	Prints information about the model
unreg_fit(y, logitx, control)	Fitting unregularized foehnix Mixture Model with Concomitant Model.