timeSerieAnalysis {FRESA.CAD}

Fit the listed time series variables to a given model

Description

This function plots the time evolution and does a longitudinal analysis of time dependent features. Features listed are fitted to the provided time model (mixed effect model) with a generalized least squares (GLS) procedure. As output, it returns the coefficients, standard errors, t-values, and corresponding p-values.

Usage

	timeSerieAnalysis(variableList,
	                  baseModel,
	                  data,
	                  timevar = "time",
	                  contime = ".",
	                  Outcome = ".",
	                  ...,
	                  description = ".",
	                  Ptoshow = c(1),
	                  plegend = c("p"),
	                  timesign = "-",
	                  catgo.names = c("Control", "Case")
	                  )

Arguments

variableList

A data frame with two columns. The first one must have the names of the candidate variables and the other one the description of such variables

baseModel

A string of the type "1 + var1 + var2" that defines the model to which variables will be fitted

data

A data frame where all variables are stored in different columns

timevar

The name of the column in data that stores the visit ID

contime

The name of the column in data that stores the continuous time (e.g. days or months) that has elapsed since the baseline visit

Outcome

The name of the column in data that stores an optional binary outcome that may be used to show the stratified analysis

description

The name of the column in variableList that stores the variable description

Ptoshow

Index of the p-values to be shown in the plot

plegend

Legend of the p-values to be shown in the plot

timesign

The direction of the arrow of time

catgo.names

The legends of the binary categories

...

Additional parameters to be passed to the gls function

Details

This function will plot the evolution of the mean value of the listed variables with its corresponding error bars. Then, it will fit the data to the provided time model with a GLS procedure and it will plot the fitted values. If a binary variable was provided, the plots will contain the case and control data. As output, the function will return the model coefficients and their corresponding t-values, and the standard errors and their associated p-values.

Value

coef

A matrix with the coefficients of the GLS fitting

std.Errors

A matrix with the standardized error of each coefficient

t.values

A matrix with the t-value of each coefficient

p.values

A matrix with the p-value of each coefficient

sigmas

The root-mean-square error of the fitting

Author(s)

Jose G. Tamez-Pena and Antonio Martinez-Torteya