version 0.1.0

DESCRIPTION

@@ -0,0 +1,21 @@
+Package: imputeTestbench
+Type: Package
+Title: Test Bench for Missing Data Imputing Models/Methods Comparison
+Version: 0.1.0
+Author: Neeraj Bokde
+Maintainer: Neeraj Bokde <neerajdhanraj@gmail.com>
+Description: Provides a Test bench for comparison of missing
+    data imputation models/methods. It compares imputing methods with reference
+    to RMSE, MAE or MAPE parameters. It allows to add new proposed methods to test
+    bench and to compare with other methods. The function 'append_method()' allows
+    to add multiple numbers of methods to the existing methods available in test bench.
+Imports: ggplot2, imputeTS, reshape2, methods, stats
+License: GPL
+LazyData: TRUE
+RoxygenNote: 5.0.1
+Suggests: knitr, rmarkdown
+VignetteBuilder: knitr
+NeedsCompilation: no
+Packaged: 2016-04-17 17:32:07 UTC; NEERAJ
+Repository: CRAN
+Date/Publication: 2016-04-17 20:11:34

MD5

@@ -0,0 +1,17 @@
+4f933b233ec56ae0ee6c747c4e928d62 *DESCRIPTION
+1ea6222a7478926f5db7d96450091059 *NAMESPACE
+c173f572f6d5ce3ccba4b8be023eb5f7 *R/append_method.R
+d669d33da67085fe3b536627fcd7a9af *R/error_functions.R
+401b4bfc21071a32d0960e1b495f31c8 *R/impute_errors.R
+803029808c7a3361ca3a5fe5d4a7473b *R/plot_errors.R
+5c37869e953de289f33b5717a3d4fdfa *build/vignette.rds
+155d7f30e87b1781bb59af8da7247618 *inst/doc/imputeTestbench-Vignette.R
+c850a692fde45c9f55f45442f1f01339 *inst/doc/imputeTestbench-Vignette.Rmd
+de0ada52dc4e87abd4a55527a404daba *inst/doc/imputeTestbench-Vignette.html
+69c68009f9e2f37fd5270640b5c4a4cb *man/append_method.Rd
+29fc1b5a9445b9947a4274a508667494 *man/impute_errors.Rd
+a1b43f0d6f7fed5e730b2234eb58d0b3 *man/mae.Rd
+0d6cdfb32eb78266af7dc3eec091c297 *man/mape.Rd
+2310169b56c8ff9707f1e6e69b89e34e *man/plot_errors.Rd
+472e93fa5117100e12f5bb00358cb759 *man/rmse.Rd
+c850a692fde45c9f55f45442f1f01339 *vignettes/imputeTestbench-Vignette.Rmd

NAMESPACE

@@ -0,0 +1,13 @@
+# Generated by roxygen2: do not edit by hand
+
+export(append_method)
+export(impute_errors)
+export(mae)
+export(mape)
+export(plot_errors)
+export(rmse)
+import(ggplot2)
+import(imputeTS)
+importFrom(methods,hasArg)
+importFrom(reshape2,melt)
+importFrom(stats,ts)

R/append_method.R

@@ -0,0 +1,138 @@
+#' To attach and compare new method to existing comparison study done with function 'impute_errors()'
+#'
+#' @param existing_method as output obtained from impute_error() function
+#' @param dataIn as imput time series for testing
+#' @param  missPercentFrom as variable from which percent of missing values to be considered
+#' @param  missPercentTo as variable to state upto what percent missing values are to be considered
+#' @param  interval as interval between consecutive missPercent values
+#' @param  repetition as an integer to state number of repetition to be done for each missPercent value
+#' @param  errorParameter as type of error calculation (RMSE, MAE or MAPE)
+#' @param  MethodPath as location of function for proposed imputation method
+#' @param  MethodName as name for function for proposed imputation method
+#' @import ggplot2
+#' @import imputeTS
+#' @importFrom methods hasArg
+#' @return Returns error comparosin for imputation methods
+#' @export
+#' @examples
+#' #Kindly, refer "Vignette" document
+
+
+#==================================================================================
+# append_method starts here....
+#==================================================================================
+
+append_method <- function(existing_method, dataIn, missPercentFrom, missPercentTo, interval, repetition, errorParameter, MethodPath, MethodName)
+{
+
+  if(!(hasArg(dataIn)))
+  {
+    dataIn <- c(1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5,1:5)
+  }
+
+  if(!is.vector(dataIn))
+  {
+    dataIn <- dataIn[, 1]
+  }
+  # For future reference
+  dataIn1 <- dataIn
+
+  # Set default values
+  if(!(hasArg(errorParameter)))
+  {
+    errorParameter <- 1
+  }
+
+  if(!(hasArg(repetition)))
+  {
+    repetition <- 1
+  }
+
+
+  if(!(hasArg(MethodName)))
+  {
+    MethodName <- "Proposed Method"
+  }
+
+  if(!(hasArg(interval)))
+  {
+    interval <- 10
+  }
+
+  if(!(hasArg(missPercentFrom)))
+  {
+    missPercentFrom <- 10
+  }
+
+  if(!(hasArg(missPercentTo)))
+  {
+    missPercentTo <- 80
+  }
+
+  e <- 0
+  f <- 0
+  # Function to create missing values
+  for(x in seq(missPercentFrom, missPercentTo, interval))
+  {
+    x <- x/100
+    a <- length(dataIn)
+    b <- a * x
+    b <- abs(b)
+    c <- a-b
+    out <- NULL
+
+    for(i in 1:repetition)
+    {
+      dataIn <- dataIn1
+      dataIn[c:(c+b)] <- NA
+      c <- sample(1:a, 1, replace = TRUE)
+      while(c > a-b)
+      {
+        c <- sample(1:a, 1, replace = TRUE)
+      }
+      out[i] <- data.frame(dataIn)
+    }
+
+    gh <- NULL
+    for(i in 1:repetition)
+    {
+      outs <- as.numeric(unlist(out[i]))
+
+      # to call functions from provided "MethodPath"
+        d <- parse(text = MethodPath)
+        d <- eval(d)
+        d <- d$value(outs)
+
+        if(errorParameter == 1)
+        {
+          gh[i] <- rmse(dataIn1 - d)
+          parameter <- "RMSE Plot"
+        }
+        if(errorParameter == 2)
+        {
+          gh[i] <- mae(dataIn1 - d)
+          parameter <- "MAE Plot"
+        }
+        if(errorParameter == 3)
+        {
+          gh[i] <- mape((dataIn1 - d), dataIn1)
+          parameter <- "MAPE Plot"
+        }
+    }
+
+
+
+      e <- append(e,mean(gh))
+      f <- append(f,x)
+  }
+    ex <- NULL
+    fx <- NULL
+    ex <- e[-1]
+    fx <- f[-1]
+    g <- data.frame(fx,ex)
+
+  #existing_method[length(existing_method)+1] <- data.frame(e[-1])
+  existing_method[[paste(MethodName)]] <- e[-1]
+  #existing_method[length(existing_method)+1] <- e
+  return(Proposed_Method = existing_method)
+}

R/error_functions.R

@@ -0,0 +1,62 @@
+#' Root Mean Square Error Calculation
+#'
+#' takes difference between Original data and Predicted data as input
+#' @param diff is the difference between Original data and Predicted data
+#' @return rmseVal as Root Mean Square Error
+#' @export
+#' @examples
+#' ## Generate 100 random numbers within some limits
+#' x <- sample(1:7, 100, replace = TRUE)
+#' y <- sample(1:4, 100, replace = TRUE)
+#' z <- rmse(x - y)
+#' z
+### RSME/MAE errors ============================================================================
+# Function that returns Root Mean Squared Error
+rmse <- function(diff)
+{
+  rmseVal <- sqrt(mean(diff^2))
+  return(rmseVal)
+}
+
+
+#' Mean Absolute Error Calculation
+#'
+#' takes difference between Original data and Predicted data as input
+#' @param diff is the difference between Original data and Predicted data
+#' @return maeVal as Mean Absolute Error
+#' @export
+#' @examples
+#' ## Generate 100 random numbers within some limits
+#' x <- sample(1:7, 100, replace = TRUE)
+#' y <- sample(1:4, 100, replace = TRUE)
+#' z <- mae(x - y)
+#' z
+# Function that returns Mean Absolute Error
+mae <- function(diff)
+{
+  maeVal <- mean(abs(diff))
+  return(maeVal)
+}
+
+
+#' Mean Absolute Percent Error Calculation
+#'
+#' takes difference between Original data and Predicted data as input
+#' @param diff is the difference between Original data and Predicted data
+#' @param dataIn as original input data
+#' @return mapeVal as Mean Absolute Error
+#' @export
+#' @examples
+#' ## Generate 100 random numbers within some limits
+#' x <- sample(1:7, 100, replace = TRUE)
+#' y <- sample(1:4, 100, replace = TRUE)
+#' z <- mape((x - y),x)
+#' z
+# Function that returns Mean Absolute Error
+mape <- function(diff, dataIn)
+{
+  mapeVal <- mean(abs(diff)* 100/dataIn)
+  return(mapeVal)
+}
+
+#==============================================================================================