Publications by T. Moudiki
Gradient-Boosting anything (alert: high performance): Part2, R version
Last week, I presented a functionality from Python package called mlsauce that allows gradient boosting of any regression algorithm. This post is about the R version. I think (?) I finally wrapped my head around the process of creating an R package from a Python package systematically, using reticulate. By default when onload ing, reticulate create...
1221 sym R (13188 sym/1 pcs) 2 img
Gradient-Boosting anything (alert: high performance)
We’ve always been told that decision trees are best for Gradient Boosting Machine Learning. I’ve always wanted to see for myself. AdaBoostClassifier is working well, but is relatively slow (by my own standards). A few days ago, I noticed that my Cython implementation of LSBoost in Python package mlsauce was already quite generic (never noticed ...
1732 sym Python (1899 sym/5 pcs) 4 img 3 tbl
Benchmarking 30 statistical/Machine Learning models on the VN1 Forecasting — Accuracy challenge
This post is about the VN1 Forecasting – Accuracy challenge. The aim is to accurately forecast future sales for various products across different clients and warehouses, using historical sales and pricing data.Phase 1 was a warmup to get an idea of what works and what wouldn’t (and… for overfitting the validation set, so that the leaderboard ...
1917 sym Python (5613 sym/19 pcs) 8 tbl
Automated random variable distribution inference using Kullback-Leibler divergence and simulating best-fitting distribution
Another post from R package misc! This time, we’ll see how to fit multiple continuous parametric distributions on a vector of data and simulate best-fitting distribution. Under the hood, misc::fit_param_dist uses a loop of MASS::fitdistr calls and Kullback-Leibler divergence for checking distribution adequacy. remotes::install_github("thierrymoud...
975 sym Python (2022 sym/5 pcs) 8 img
Forecasting in Excel using Techtonique’s Machine Learning APIs under the hood
I created a basic Python FastAPI app that allows you to interact with Techtonique web app through Excel. More specifically, Visual Basic for Applications (VBA) Excel. All you need to do is download the Excel file VBA-Web.xlsm, visit “Sheet1” and “Sheet2”, and click the buttons. In “Sheet2”, you should see something like this:Keep in min...
3886 sym 1 img
Techtonique web app for data-driven decisions using Mathematics, Statistics, Machine Learning, and Data Visualization
This week, I released Techtonique web app, a tool designed to help you make informed, data-driven decisions using Mathematics, Statistics, Machine Learning, and Data Visualization. As of September 2024, the tool is in its beta phase (subject to crashes) and will remain completely free to use until December 24, 2024. After registering, you will rece...
2982 sym 1 img 5 tbl
Parallel for loops (Map or Reduce) + New versions of nnetsauce and ahead
nnetsauceThe news are (reminder: the nnetsauce.Lazy*s do automated Machine Learning benchmarking of multiple models):Update LazyDeepMTS: no more LazyMTS class, instead, you can use LazyDeepMTS with n_layers=1Specify forecasting horizon in LazyDeepMTS (see updated docs and examples/lazy_mts_horizon.py)New class ClassicalMTS for classsical models (fo...
1707 sym R (655 sym/10 pcs)
Adaptive (online/streaming) learning with uncertainty quantification using Polyak averaging in learningmachine
The model presented here is a frequentist – conformalized – version of the Bayesian one presented in #152. It is implemented in learningmachine, both in Python and R, and is updated as new observations arrive, using Polyak averaging. Model explanations are given as sensitivity analyses. 1 – R version %load_ext rpy2.ipython %%R utils::instal...
828 sym R (13827 sym/6 pcs) 6 img 4 tbl
Adaptive (online/streaming) learning with uncertainty quantification using Polyak averaging in learningmachine
The model presented here is a frequentist – conformalized – version of the Bayesian one presented in #152. It is implemented in learningmachine, both in Python and R, and is updated as new observations arrive, using Polyak averaging. Model explanations are given as sensitivity analyses.1 – R version%load_ext rpy2.ipython %%R utils::install.p...
923 sym R (13786 sym/28 pcs) 3 img 4 tbl
New versions of nnetsauce and ahead
New versions of nnetsauce and ahead will be released this week. Big releases, notably with more uncertainty quantification methods implemented in both packages. Stay tuned. PS: You can now (well, again) comment on posts on this website. Hope it works better than Disqus. Related To leave a comment for the author, please follow the link and comment...
676 sym