README.md

mephas

MEPHAS is a shiny-based statistical software that was developed to support statistical data analyses for medical and pharmaceutical students, researchers, and doctors.

Two platforms are available: (1) web server and (2) R package.

Web server MEPHAS

Homepage and source code at GitHub

R package mephas, installations

We suggested to use the latest version of R and RStudio. The old version may not support some visualization functions

R package needs to be installed in R console. If you have not installed R, download R first:

You can use RStudio console, R console, and even command line interface to open the R environment

The followings show how to install in RStudio, which is the same in R console and R in command line interface

Installation from GitHub in RStudio console

To install package from GitHub, we need R package "remotes". Package "mephas" need to use some functions from "mephas.tools", so we need to install 2 packages.

First, check the existence of R package "remotes"

> packageVersion("remotes")

If error message said "there is no package called 'remotes', then install "remotes" package first.

> install.packages("remotes")

If R package "remotes" has been installed, then install "mephas.tools" and "mephas" package from GitHub

> remotes::install_github(c("mephas/mephas.tools", "mephas/mephas"), upgrade="never")

Note: upgrade="never": suppress the updates of the R packages, see details here; package "remotes" and "devtools" shared the same install_github function, thus package "remotes" can be replaced by package "devtools"

Videos of the installation process on windows

Installation from ".tar.gz" file

Installing from ".tar.gz" file is not recommended. Users need to install two packages: "mephas.tools" and "mephas". Some packages are also required.

Download files: mephas.tools_1.0.0.tar.gz and mephas_1.1.tar.gz (updated 2020-2)

Check whether you have installed the following packages in R
> mephas.need.packages <- c("DescTools",
                        "DT",
                        "exactRankTests",
                        "dunn.test",
                        "ROCR",
                        "ggplot2",
                        "magrittr",
                        "psych",
                        "pls",
                        "plotly",
                        "reshape",
                        "shiny",
                        "shinythemes",
                        "shinyWidgets",
                        "survival",
                        "survminer",
                        "survAUC",
                        "spls")
> not.installed.packages <- mephas.need.packages[!(mephas.need.packages %in% installed.packages()[,"Package"])]
> not.installed.packages
If some packages have not been installed, install them first
> if(length(not.installed.packages)) install.packages(not.installed.packages)

Videos of the installation process on windows

Find a statistical methods using the flowchart

Flowchart.pdf

Commands to open mephas interfaces

Load the package

Before open the applications, we need to load the package "mephas" together with "mephas.tools"

> library(mephas)

Then, you can open the applications as follows.

All-in-one function, see details

> mephasOpen()

Alternatively, open each interfaces using the method names

1. Probability distributions

Continuous probability distribution, web server

> MFScondist()

Discrete probability distribution, web server

> MFSdisdist()
2. Parametric T test for means, web server
> MFSttest()
3. Non-parametric tests for median, web server
> MFSnptest()
4. Test for binomial proportions, web server
> MFSproptest()
5. Test for contingency tables, web server
> MFSrctabtest()
6. Analysis of variance, web server
> MFSanova()
7. Statistical model,

Linear regression, web server

> MFSlr()

Logistic regression, web server

> MFSlogit()

Survival analysis, web server

> MFSsurv()
8. Dimensional analysis

Dimensional analysis 1, web server

> MFSpca()

Dimensional analysis 2, web server

> MFSpls()

More functions are under construction...



mephas/mephas documentation built on Feb. 20, 2020, 9:03 p.m.