morpheus.git
descriptionSpectral methods to estimate mixture parameters
last changeMon, 6 Nov 2023 20:17:40 +0000 (21:17 +0100)

Estimate parameters of mixtures of logistic regressions

This code is one applied part in the PhD thesis of Mor-Absa Loum

Description

Mixture of lOgistic Regressions Parameters (H)Estimation with (U)Spectral methods. The main methods take d-dimensional inputs + a vector of binary outputs, and return parameters according to the GLMs mixture model (please see the package vignette).

NOTE: greek unicode letters are used in the code, because it’s much nicer to write λ, β and Σ than lambda, beta and Sigma - and also more importantly because it works well :) …However CRAN demands ASCII files - therefore the presence of to-cran.sh script.

Warning: tests won’t run well on the non-UTF8 package. Run ./to-cran.sh and then inside pkg-cran call devtools::test().

Example

Install the package, then ?multiRun is a possible starting point.

shortlog
2023-11-06 Benjamin Auderupdate master
2023-09-13 Benjamin AuderRemove OpenMP dependency (and unused variable)
2023-02-07 Benjamin AuderSome fixes to re-upload on the CRAN
2021-06-07 Benjamin AuderAlso version NAMESPACE because roxygen2 complains when...
2021-06-07 Benjamin AuderFix computeW (was random - TODO)
2021-06-07 Benjamin AuderFix accuracy.R + add postTreatment draft
2021-06-07 Benjamin AuderAdjustments + bugs fixing
2020-04-30 Benjamin AuderError when K > d for computeMu and optimParams
2020-01-15 Benjamin AuderAllow to optimize from a given matrix W
2020-01-15 Benjamin AuderAdd draft of a function to draw computation times
2020-01-14 Benjamin AuderFix typo
2020-01-14 Benjamin AuderIt's enough to compute half of W (complete by symmetry)
2020-01-14 Benjamin AuderAdd a number_of_cores parameter for OpenMP // in Comput...
2020-01-12 Benjamin AuderUpdate year in LICENSE file, use default compilation...
2020-01-10 Benjamin AuderFix Mor thesis link + compiling for Windows
2019-12-23 Benjamin AuderPackage v.1.0 ready to be sent to CRAN
...
heads
13 months ago master