R/plotFit.R

Defines functions plotFit

Documented in plotFit

## Written by Mercedeh Movassagh <mercedeh@ds.dfci.harvard.edu>, Aug 2020

#' @importFrom graphics par abline points mtext
#' @importFrom stats pf lm
NULL

#' Plot model
#'
#' Plot 2D description
#' @param model linear model
#' @return does not return value
#' @export
#' @keywords plot, 2d, model
#' @examples
#' plotFit(lm(x ~ y, data = data.frame(x = runif(10), y = runif(10))))
plotFit <- function(model) {
    m <- model$model
    m$Fitted <- model$fitted.values
    last <- ncol(m)
    names(m)[last] <- "Fitted values\nlm(model_formula)"
    par(mar = c(5, 5, 4, 2) + 0.1)
    plot(m[, c(1, last)],
        type = "n", main = as.character.formula(model$terms),
        cex.main = 1
    )
    abline(lm(m[, last] ~ m[, 1]), lwd = 2, col = "red")
    points(m[, c(1, last)])

    k <- length(model$coefficients) - 1 # Number of mirnas in model!
    n <- length(model$fitted.values)
    rs <- summary(model)$r.squared
    Pvalue <- pf((rs / k) / ((1 - rs) / (n - 1 - k)), k, n - 1 - k, lower.tail = FALSE) # p value for R-squared

    mtext(sprintf("P value: %.2g", Pvalue), line = .74)
    mtext(sprintf("R-squared: %.2f", summary(model)$r.squared), line = .01)
    return(invisible(NULL))
}
Mercedeh66/mirTarRnaSeq documentation built on April 14, 2023, 6:49 a.m.