Commit | Line | Data |
---|---|---|
3d5b5060 BA |
1 | --- |
2 | title: morpheus........... | |
3 | ||
4 | output: | |
5 | pdf_document: | |
6 | number_sections: true | |
7 | toc_depth: 1 | |
8 | --- | |
9 | ||
10 | ```{r setup, results="hide", include=FALSE} | |
11 | knitr::opts_chunk$set(echo = TRUE, include = TRUE, | |
12 | cache = TRUE, comment="", cache.lazy = FALSE, | |
13 | out.width = "100%", fig.align = "center") | |
14 | ``` | |
15 | ||
16 | 0) Tell that we try to learn classification parameters in a non-EM way, using algebric manipulations. | |
3d5b5060 | 17 | |
cff1083b BA |
18 | *morpheus* is a contributed R package which attempts to find the parameters of a mixture of logistic classifiers. |
19 | When the data under study come from several groups that have different characteristics, using mixture models is a very popular way to handle heterogeneity. | |
20 | Thus, many algorithms were developed to deal with various mixtures models. Most of them use likelihood methods or Bayesian methods that are likelihood dependent. | |
21 | *flexmix* is an R package which implements these kinds of algorithms. | |
3d5b5060 | 22 | |
cff1083b BA |
23 | However, one problem of such methods is that they can converge to local maxima, so several starting points must be explored. |
24 | Recently, spectral methods were developed to bypass EM algorithms and they were proved able to recover the directions of the regression parameter | |
25 | in models with known link function and random covariates (see [9]). | |
26 | Our package extends such moment methods using least squares to get estimators of the whole parameters (with theoretical garantees, see [XX]). | |
27 | Currently it can handle only binary output $-$ which is a common case. | |
3d5b5060 | 28 | |
cff1083b | 29 | 1) Model. |
3d5b5060 | 30 | |
cff1083b | 31 | TODO: retrouver mon texte initial + article. |
3d5b5060 | 32 | |
cff1083b | 33 | 2) Algorithm (as in article) |
3d5b5060 | 34 | |
cff1083b | 35 | 3) Experiments: show package usage |