In Jorisvdoorn/bonuslabgroup8: Introduction to Machine Learning in R

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(bonuslabgroup8)
library(caret)
library(mlbench)
library(leaps)

Simple Prediction using ridgereg()

This vignette showcases how to do a simple prediction using the ridgereg() function.

reg1 = ridgereg(Petal.Length~Species, data = iris, lambda = 0.01)
reg1$print()
head(reg1$predict())

Create predictive models

This example uses data from the BostonHousing dataset.

data("BostonHousing")

Data

The data is split into a training and test set using the caret package, with 80% of the data in the training set and 20% of the data in the test set.

set.seed(3456)
trainIndex <- createDataPartition(BostonHousing$medv, p = 0.8, list = FALSE, times = 1)
bhTrain <- BostonHousing[trainIndex,]
bhTest <- BostonHousing[-trainIndex,]

Fit Linear regression and linear regression with forward selection

lmFit = caret::train(medv ~ ., data = bhTrain, method = "lm")
lmFit

lfFit = caret::train(medv ~ ., data = bhTrain, method = "leapForward")
lfFit

By looking at the metrics (RMSE, Rsquared, and MAE), we can see that the linear regression performs better than the linear regression with forward selection.

Fit a ridge regression model

rr_create = function(){
  rr = list(type = "Regression", library = "bonuslabgroup8", loop = NULL, prob = NULL)
  prm = data.frame(parameter = "lambda", class = "numeric", label = "Lambda")
  rr$parameters = prm
  rrgrid = function(x, y, len = NULL, search = "grid"){
    data.frame(lambda = c(0.5, 0.1, 0.05, 0.01, 0.005, 0.001))
  }
  rr$grid = rrgrid
  rrfit <- function(x, y, wts, param, lev, last, weights, classProbs, ...) { 
    df = as.data.frame(x)
    df$resp = y
    ridgereg(formula = resp ~ ., data = df, lambda = param$lambda, ...)
  }
  rr$fit = rrfit
  rrpred <- function(modelFit, newdata, preProc = NULL, submodels = NULL){
    modelFit$predict(newdata)
  }
  rr$predict <- rrpred
  rrsort <- function(x) x[order(x$lambda),]
  rr$sort <- rrsort
  rr$levels <- function(x) x@levels
  return(rr)
}

rr_method = rr_create()

set.seed(998)
fitControl <- trainControl(method = "repeatedcv",
                           number = 10,
                           repeats = 10)

set.seed(825)
ridgeFit <- train(medv ~.,
                  data = bhTrain,
                  method = rr_method,
                  trControl = fitControl
                     )


# keep getting errors: predictions failed for ... Error in modelFit$predict(newdata) : unused argument (newdata)

Jorisvdoorn/bonuslabgroup8 documentation built on Oct. 30, 2019, 8:01 p.m.