NNSummary: Summarize Calculations of RMSE, MSE, MAE, and WAE

In pkR-pkR/NNbenchmark: Datasets and Functions to Benchmark Neural Network Packages

Summarize Calculations of RMSE, MSE, MAE, and WAE

Description

Summarize measures of fit and time for a single training. Measures of fit include the Root Mean Squared Error (RMSE), the Mean Squared Error (MSE), the Mean Absolute Error (MAE), and the Worst Absolute Error (WAE) rounded by default to 4 digits and set to na.rm = TRUE. See more at funRMSE. The summary can also include the results of time from getTimer in NNbenchmark or the result of timediff.

Usage

NNsummary(y_pred, y0, time, dgts = 4)

Arguments

y_pred	numeric vector of the predicted values
y0	numeric vector of the observed values
time	numeric value of time
dgts	integer value for how many digits to round to, see round

Value

A vector of RMSE, MSE, MAE, WAE, and time values for a single iteration.

Examples

## With 2019 legacy code, no longer usable with 2020 trainPredict
old <- options("digits.secs" = 4)
timeTT <- createTimer()

timeTT$start("event")
y0 <- 1:19
y_pred <- y0 + rnorm(length(y0), sd = 0.3)
timeTT$stop("event") 

time <- getTimer(timeTT)

NNsummary(y_pred, y0, time[,4], 4)

## With 2020 code
timestart()
y0 <- 1:19
y_pred <- y0 + rnorm(length(y0), sd = 0.3)
time <- timediff()

NNsummary(y_pred, y0, time, 4)

options(old)      
  

pkR-pkR/NNbenchmark documentation built on Dec. 3, 2023, 12:43 a.m.