fit_params

Functions to estimate parameters of probability distributions by fitting the distributions using optim()

Estimation

Parameter

Provides functions for fitting discrete distribution models to count data.
Included are the Poisson, the negative binomial, the Poisson-inverse gaussian and, most importantly,
a new implementation of the Poisson-beta distribution (density, distribution and quantile
functions, and random number generator) together with a needed new implementation of
Kummer's function (also: confluent hypergeometric function of the first kind). Three
different implementations of the Gillespie algorithm allow data simulation based on the
basic, switching or bursting mRNA generating processes. Moreover, likelihood functions for
four variants of each of the three aforementioned distributions are also available.
The variants include one population and two population mixtures, both with and without
zero-inflation. The package depends on the 'MPFR' libraries (<https://www.mpfr.org/>) which need to be installed separately
(see description at <https://github.com/fuchslab/scModels>).
This package is supplement to the paper "A mechanistic model for the negative binomial distribution of single-cell mRNA counts"
by Lisa Amrhein, Kumar Harsha and Christiane Fuchs (2019) <doi:10.1101/657619> available on bioRxiv.

Lisa Amrhein

scModels

Fitting Discrete Distribution Models to Count Data

Kumar Harsha

Christiane Fuchs

Pavel Holoborodko

fit_params function

<dl><dt>x</dt>
<dd>Vector containing the discrete observations</dd>
<dt>type</dt>
<dd>Keyword for the probability distribution the data is to be fitted
against. Possible values are ("pois", "nb", "del", pig", "pb", "pois2", "nb2", "del2", "pig2"
 "pb2", "zipois", "zinb", "zidel", "zipg", zipb", "zipois2", "zinb2", "zidel2", "zipig2", zipb2")</dd>
<dt>optim_control</dt>
<dd>List of options to override presets in
the optim function; Set to list(maxit = 1000) by default.
For more details, please refer to the 'control' parameter in the
standard 'optim' function in package 'stats'.</dd></dl>

Arguments

Functions to estimate parameters of probability distributions by fitting the distributions using optim() — fit_params

<dl>

<dt>x</dt>
<dd>Vector containing the discrete observations</dd>


<dt>type</dt>
<dd>Keyword for the probability distribution the data is to be fitted
against. Possible values are ("pois", "nb", "del", pig", "pb", "pois2", "nb2", "del2", "pig2"
 "pb2", "zipois", "zinb", "zidel", "zipg", zipb", "zipois2", "zinb2", "zidel2", "zipig2", zipb2")</dd>


<dt>optim_control</dt>
<dd>List of options to override presets in
the optim function; Set to list(maxit = 1000) by default.
For more details, please refer to the 'control' parameter in the
standard 'optim' function in package 'stats'.</dd>

</dl>

Last chance! 50% off unlimited learning

fit_params: Functions to estimate parameters of probability distributions by fitting the distributions using optim()

Description

Usage

Arguments

Examples