Publications by T. Moudiki
learningmachine v1.1.2: for Python
I talked about learningmachine – a package for machine learning with uncertainty quantification and interpretatbility – last week in #131. Here comes the Python version!Keep in mind that learningmachine is still experimental, probably with some quirks (because achieving this level of abstraction required some effort), with no beautiful document...
698 sym Python (7492 sym/15 pcs) 2 img
Bayesian inference and conformal prediction (prediction intervals) in nnetsauce v0.18.1
Version 0.18.1 of nnetsauce (Python version) is available on PyPI and for conda. New developments include Bayesian inference and conformal prediction. Bayesian inference is available for scikit-learn models that possess a posterior distribution (BayesianRidge, ARDRegressor, and GaussianProcessRegressor). Conformal prediction is available for every ...
943 sym Python (4739 sym/16 pcs) 5 img
rtopy (v0.1.1): calling R functions in Python
rtopy is a Python package that allows you to call R functions in Python. There are other packages doing something similar, but in addition to have a different interface, under the hood, rtopy explicitly uses R at the command line, plus text mining and caching tools. It’s a work in progress, and you can find some examples of use below.1 – Instal...
573 sym Python (4976 sym/17 pcs) 1 img
ahead forecasting (v0.10.0): fast time series model calibration and Python plots
ahead is an R, Python and Julia package for univariate and multivariate time series forecasting with uncertainty quantification (including predictive simulation). Matlab is next. The aim is always to make these implementations of ahead as similar as possible, but that’s a looot of work as you might guess. Do not hesitate to submit a pull request...
3243 sym R (3187 sym/22 pcs) 6 img 1 tbl
Subsampling continuous and discrete response variables
New in nnetsauce v0.16.3:add robust scaler to type_scalingrelatively faster scaling in preprocessingRegression-based classifiers (see https://www.researchgate.net/publication/377227280_Regression-based_machine_learning_classifiers)DeepMTS (multivariate time series forecasting with deep quasi-random layers): see https://thierrymoudiki.github.io/blog...
1048 sym 3 img
Tuning Machine Learning models with GPopt’s new version
A new version of Python package GPopt is available on PyPI. GPopt is a package for stochastic optimization based on Gaussian process regressors (for now, the name GP* is ‘unfortunate’). This type of optimization is particularly useful for tuning machine learning models’ hyperparameters.The main change in GPopt’s v0.3.0 is: the user can now ...
760 sym 3 img
nnetsauce’s introduction as of 2024-02-11 (new version 0.17.0)
0 – What is nnetsauce?nnetsauce is a general-purpose computational tool for statistical/Machine Learning (ML). As of 2024-02-11, nnetsauce is available for R and Python. It contains various implementations of – shallow and deep – regression, classification, univariate and multivariate probabilistic time series forecasting models.The specifici...
4777 sym Python (9365 sym/21 pcs) 1 img
Tuning Machine Learning models with GPopt’s new version Part 2
A new version of Python package GPopt is available on PyPI. GPopt is a package for stochastic optimization. This type of optimization is particularly useful for tuning machine learning models’ hyperparameters.The main change in GPopt’s v0.4.0 is: the user can name the objective function’s parameters. This makes it easier to use the results di...
727 sym Python (4022 sym/12 pcs) 1 img
A plethora of datasets at your fingertips Part3: how many times do couples cheat on each other?
Remark: I have no affiliation with these projects, but I love using quarto and Visual Studio Code’s extension Jupyter To Markdown for these posts. Just wanted to say that.Today:I’ll show you how to use Python’s nnetsauce (version 0.17.0) to download data from the R universe API.I’ll use nnetsauce’s automated Machine Learning (AutoML) to a...
1013 sym Python (3910 sym/13 pcs)
learningmachine: prediction intervals for conformalized Kernel ridge regression and Random Forest
I created R package learningmachine in the first place, in order to have a unified, object-oriented (using R6), interface for the machine learning algorithms I use the most on tabular data. This (work in progress) package is available on GitHub and the R-universe. There will certainly be a Python version in the future. This post shows how to use le...
1023 sym R (5065 sym/5 pcs) 2 img