Prediction Assignment Writeup
Ing. Greg Bennett Gibb
March 6, 2018
1. Background
Using devices such as Jawbone Up, Nike FuelBand, and Fitbit it is now possible to collect a large amount of data about personal activity relatively inexpensively. These type of devices are part of the quantified self movement - a group of enthusiasts who take measurements about themselves regularly to improve their health, to find patterns in their behavior, or because they are tech geeks. One thing that people regularly do is quantify how much of a particular activity they do, but they rarely quantify how well they do it. In this project, your goal will be to use data from accelerometers on the belt, forearm, arm, and dumbell of 6 participants. They were asked to perform barbell lifts correctly and incorrectly in 5 different ways. More information is available from the website here: http://groupware.les.inf.puc-rio.br/har (see the section on the Weight Lifting Exercise Dataset). Knit ### 2. Data
The training data for this project are available here: https://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv
The test data are available here:
https://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv
library(caret); library(rpart);library(ggplot2);library(randomForest)## Warning: package 'caret' was built under R version 3.4.3## Loading required package: lattice## Loading required package: ggplot2## Warning: package 'ggplot2' was built under R version 3.4.3## Warning: package 'rpart' was built under R version 3.4.3## Warning: package 'randomForest' was built under R version 3.4.3## randomForest 4.6-12## Type rfNews() to see new features/changes/bug fixes.##
## Attaching package: 'randomForest'## The following object is masked from 'package:ggplot2':
##
## margin2.1 Loading Data
CHeck the training and testing data, identifying the missing data, “NA” and “#DIV/0!” as “NA” everywhere.
url.train <- "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv"
url.test <- "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv"
training <- read.csv(url(url.train), na.strings = c("NA", "", "#DIV0!"))
testing <- read.csv(url(url.test), na.strings = c("NA", "", "#DIV0!"))We need to define the same columns
sameColumsName <- colnames(training) == colnames(testing)
colnames(training)[sameColumsName==FALSE]## [1] "classe"Therefore,the “classe” is not included in the testing data. #### 2.2 Cleaning Data
training<-training[,colSums(is.na(training)) == 0]
testing <-testing[,colSums(is.na(testing)) == 0]2.3 Checking the column names of traning dataset
You can also embed plots, for example:
head(colnames(training))## [1] "X" "user_name" "raw_timestamp_part_1"
## [4] "raw_timestamp_part_2" "cvtd_timestamp" "new_window"The first 7 variables of the training data were deleted, because they are irrelevant to the prediction. #### 2.4.Training, testing & validation data The training dataset was separated into three parts: tranining part (60%), testing part (20%), and validation part (20%)
set.seed(123)
Seeddata1 <- createDataPartition(y = training$classe, p = 0.8, list = F)
Seeddata2 <- training[Seeddata1,]
validation <- training[-Seeddata1,]
Training_data1 <- createDataPartition(y = Seeddata2$classe, p = 0.75, list = F)
training_data2 <- Seeddata2[Training_data1,]
testing_data <- Seeddata2[-Training_data1,]
qplot(classe, fill = "4", data=training_data2, main="Distribution of Classes")
names(training_data2[,-53])## [1] "X" "user_name" "raw_timestamp_part_1"
## [4] "raw_timestamp_part_2" "cvtd_timestamp" "new_window"
## [7] "num_window" "roll_belt" "pitch_belt"
## [10] "yaw_belt" "total_accel_belt" "gyros_belt_x"
## [13] "gyros_belt_y" "gyros_belt_z" "accel_belt_x"
## [16] "accel_belt_y" "accel_belt_z" "magnet_belt_x"
## [19] "magnet_belt_y" "magnet_belt_z" "roll_arm"
## [22] "pitch_arm" "yaw_arm" "total_accel_arm"
## [25] "gyros_arm_x" "gyros_arm_y" "gyros_arm_z"
## [28] "accel_arm_x" "accel_arm_y" "accel_arm_z"
## [31] "magnet_arm_x" "magnet_arm_y" "magnet_arm_z"
## [34] "roll_dumbbell" "pitch_dumbbell" "yaw_dumbbell"
## [37] "total_accel_dumbbell" "gyros_dumbbell_x" "gyros_dumbbell_y"
## [40] "gyros_dumbbell_z" "accel_dumbbell_x" "accel_dumbbell_y"
## [43] "accel_dumbbell_z" "magnet_dumbbell_x" "magnet_dumbbell_y"
## [46] "magnet_dumbbell_z" "roll_forearm" "pitch_forearm"
## [49] "yaw_forearm" "total_accel_forearm" "gyros_forearm_x"
## [52] "gyros_forearm_y" "accel_forearm_x" "accel_forearm_y"
## [55] "accel_forearm_z" "magnet_forearm_x" "magnet_forearm_y"
## [58] "magnet_forearm_z" "classe"3.Prediction model (Classification Tree model)
model_tree <- rpart(classe ~ ., data=training_data2, method="class")
prediction_tree <- predict(model_tree, testing_data, type="class")
class_tree <- confusionMatrix(prediction_tree, testing_data$classe)
class_tree## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1116 0 0 0 0
## B 0 759 0 0 0
## C 0 0 684 0 0
## D 0 0 0 643 0
## E 0 0 0 0 721
##
## Overall Statistics
##
## Accuracy : 1
## 95% CI : (0.9991, 1)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 1
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 1.0000 1.0000 1.0000 1.0000 1.0000
## Specificity 1.0000 1.0000 1.0000 1.0000 1.0000
## Pos Pred Value 1.0000 1.0000 1.0000 1.0000 1.0000
## Neg Pred Value 1.0000 1.0000 1.0000 1.0000 1.0000
## Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Detection Rate 0.2845 0.1935 0.1744 0.1639 0.1838
## Detection Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Balanced Accuracy 1.0000 1.0000 1.0000 1.0000 1.00003.1.Checking the model_tree
library(rpart.plot)## Warning: package 'rpart.plot' was built under R version 3.4.3rpart.plot(model_tree)
3.2.Random forest model
forest_model <- randomForest(classe ~ ., data=training_data2, method="class")
prediction_forest <- predict(forest_model, testing_data, type="class")
random_forest <- confusionMatrix(prediction_forest, testing_data$classe)
random_forest## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1116 0 0 0 0
## B 0 759 0 0 0
## C 0 0 684 0 0
## D 0 0 0 643 0
## E 0 0 0 0 721
##
## Overall Statistics
##
## Accuracy : 1
## 95% CI : (0.9991, 1)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 1
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 1.0000 1.0000 1.0000 1.0000 1.0000
## Specificity 1.0000 1.0000 1.0000 1.0000 1.0000
## Pos Pred Value 1.0000 1.0000 1.0000 1.0000 1.0000
## Neg Pred Value 1.0000 1.0000 1.0000 1.0000 1.0000
## Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Detection Rate 0.2845 0.1935 0.1744 0.1639 0.1838
## Detection Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Balanced Accuracy 1.0000 1.0000 1.0000 1.0000 1.00003.3 .Final prediction
Prediction Algorithm and Confusion Matrix
prediction1 <- predict(forest_model, newdata=testing_data)
confusionMatrix(prediction1, testing_data$classe)## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1116 0 0 0 0
## B 0 759 0 0 0
## C 0 0 684 0 0
## D 0 0 0 643 0
## E 0 0 0 0 721
##
## Overall Statistics
##
## Accuracy : 1
## 95% CI : (0.9991, 1)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 1
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 1.0000 1.0000 1.0000 1.0000 1.0000
## Specificity 1.0000 1.0000 1.0000 1.0000 1.0000
## Pos Pred Value 1.0000 1.0000 1.0000 1.0000 1.0000
## Neg Pred Value 1.0000 1.0000 1.0000 1.0000 1.0000
## Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Detection Rate 0.2845 0.1935 0.1744 0.1639 0.1838
## Detection Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Balanced Accuracy 1.0000 1.0000 1.0000 1.0000 1.0000The Random Forest is a much better predictive model than the Decision Tree, which has a larger accuracy (99.91%).Therefore, we don’t need to consider more important predictors for the Random Forest model 4.Conclusions In this study, the characteristics of predictors for both traning and testing datasets (train and test) are reduced. These characteristics are the percentage of NAs values, low variance, correlation and skewness. Therefore, the variables of the data sets are scaled. The training dataset is splitted into subtraining and validation parts to construct a predictive model and evaluate its accuracy. Decision Tree and Random Forest are applied.The Random Forest is a much better predictive model than the Decision Tree, which has a larger accuracy (99.91%). This project is reproducible and was done with the following environment:
sessionInfo()## R version 3.4.1 (2017-06-30)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 7 x64 (build 7601) Service Pack 1
##
## Matrix products: default
##
## locale:
## [1] LC_COLLATE=English_United States.1252
## [2] LC_CTYPE=English_United States.1252
## [3] LC_MONETARY=English_United States.1252
## [4] LC_NUMERIC=C
## [5] LC_TIME=English_United States.1252
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] rpart.plot_2.1.2 randomForest_4.6-12 rpart_4.1-13
## [4] caret_6.0-78 ggplot2_2.2.1 lattice_0.20-35
##
## loaded via a namespace (and not attached):
## [1] Rcpp_0.12.12 lubridate_1.6.0 tidyr_0.7.1
## [4] class_7.3-14 assertthat_0.2.0 rprojroot_1.3-2
## [7] digest_0.6.15 ipred_0.9-6 psych_1.7.8
## [10] foreach_1.4.4 R6_2.2.2 plyr_1.8.4
## [13] backports_1.1.0 stats4_3.4.1 e1071_1.6-8
## [16] evaluate_0.10.1 rlang_0.1.2 lazyeval_0.2.0
## [19] kernlab_0.9-25 Matrix_1.2-10 rmarkdown_1.9
## [22] labeling_0.3 splines_3.4.1 CVST_0.2-1
## [25] ddalpha_1.3.1.1 gower_0.1.2 stringr_1.2.0
## [28] foreign_0.8-69 munsell_0.4.3 broom_0.4.3
## [31] compiler_3.4.1 pkgconfig_2.0.1 mnormt_1.5-5
## [34] dimRed_0.1.0 htmltools_0.3.6 nnet_7.3-12
## [37] tidyselect_0.2.0 tibble_1.3.4 prodlim_1.6.1
## [40] DRR_0.0.3 codetools_0.2-15 RcppRoll_0.2.2
## [43] dplyr_0.7.2 withr_2.1.1 MASS_7.3-47
## [46] recipes_0.1.2 ModelMetrics_1.1.0 grid_3.4.1
## [49] nlme_3.1-131 gtable_0.2.0 magrittr_1.5
## [52] scales_0.5.0 stringi_1.1.5 reshape2_1.4.2
## [55] bindrcpp_0.2 timeDate_3043.102 robustbase_0.92-8
## [58] lava_1.6 iterators_1.0.9 tools_3.4.1
## [61] glue_1.1.1 DEoptimR_1.0-8 purrr_0.2.3
## [64] sfsmisc_1.1-2 parallel_3.4.1 survival_2.41-3
## [67] yaml_2.1.14 colorspace_1.3-2 knitr_1.20
## [70] bindr_0.1