Michael RoweFast, Online Forecasting PackagesHere is how to use incremental, online time-series prediction algorithms or glance at their relative speed in the Elo ratings. It is based…Nov 1, 2023Nov 1, 2023
Michael RoweThe Clean Way to Deal With Ordinal Values in SkLearnOrdinal data is a little annoying, but less so with sklearn.Oct 15, 2023Oct 15, 2023
Michael RowePresidential Poetry AppreciationBecause why not? I stole this fun example from the examples directory of the chattychattybangbang package. Here are the presidents that…Sep 19, 2023Sep 19, 2023
Michael RoweHow to Use the Huber Pseudo-MeanA note on the accuracy of bootstrapped forecasts and Huber means.Sep 2, 2023Sep 2, 2023
Michael RoweCreating a Memory-Frugal Torch Dataset from a Collection of CSVsThe Python csvsdataset package does just one thing, and if you have a large collection of CSV data files that you don’t want to put in…Aug 8, 2023Aug 8, 2023
Michael RoweClustering for Efficient Computation of Contest Winning ProbabilitiesHere’s a contest with three “fast” horses and several slow ones. How shall we quickly estimate winning probabilities? Could we do it…Aug 1, 2023Aug 1, 2023
Michael RoweWhy Doesn’t PyCaret Work Better for Time-Series?PyCaret doesn’t have a very good Elo rating for univariate timeseries. It is also very computationally expensive. Worst of both worlds…Jul 23, 2023Jul 23, 2023
Michael RowePredicting Volatility Using PyKalman’s Implementation of the Expectation Maximization (EM)…I present a colab notebook example of using the Expectation Maximization algorithm to predict noisy data.Jun 16, 2023Jun 16, 2023
Michael RoweMaking a Safe, Simple Online Version of NeuralProphetNeuralProphet is a Python package for time-series forecasting that is built on the principles of Prophet, an open-source forecasting…Jun 3, 2023Jun 3, 2023
Michael RoweFractional Differencing in PyTorch, with ReconstructionGottfried Wilhelm Leibniz first suggested the possibility of fractional derivatives in the 17th Century. It allows for a nuanced approach…Jun 1, 2023Jun 1, 2023