Distributed model-fitting in Python

[sryza]

Right now it's not possible to fit an ARIMA model to time series in a distributed fashion. It could make sense to add a method to the TimeSeriesRDD class fit_model(model_name, **kwargs) where the kwargs would be fitting parameters.

Alternatively, the ARIMA module could have a fit_each function which accepts a time series RDD and fits a model to each series in it.