Practical Machine Learning Coursera Project
Tanishq Porwal
11/08/2020
Project Information:
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).
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
The data for this project come from this source: http://groupware.les.inf.puc-rio.br/har. If you use the document you create for this class for any purpose please cite them as they have been very generous in allowing their data to be used for this kind of assignment.
Goal (What you should submit)
The goal of your project is to predict the manner in which they did the exercise. This is the “classe” variable in the training set. You may use any of the other variables to predict with. You should create a report describing how you built your model, how you used cross validation, what you think the expected out of sample error is, and why you made the choices you did. You will also use your prediction model to predict 20 different test cases.
#Loading the necessary libraries:
library(caret)
library(rpart)
library(ggplot2)
library(e1071)
library(rattle)Loading Data:
#Loading the necessary libraries:
train <- read.csv("E:\\RStudio\\projects\\PracticalMachineLearning\\pml-training.csv",na.strings=c("NA","#DIV/0!",""))
test <- read.csv("E:\\RStudio\\projects\\PracticalMachineLearning\\pml-testing.csv", na.strings=c("NA","#DIV/0!",""))Partitioning the training dataset into two parts:
#Loading the necessary libraries:
set.seed(12345)
t <- createDataPartition(train$classe, p=0.6, list = FALSE)
train1 <- train[t, ]
valid <- train[-t, ]
dim(train1); dim(valid)## [1] 11776 160## [1] 7846 160head(train1)## X user_name raw_timestamp_part_1 raw_timestamp_part_2 cvtd_timestamp
## 1 1 carlitos 1323084231 788290 05/12/2011 11:23
## 2 2 carlitos 1323084231 808298 05/12/2011 11:23
## 5 5 carlitos 1323084232 196328 05/12/2011 11:23
## 9 9 carlitos 1323084232 484323 05/12/2011 11:23
## 10 10 carlitos 1323084232 484434 05/12/2011 11:23
## 14 14 carlitos 1323084232 576390 05/12/2011 11:23
## new_window num_window roll_belt pitch_belt yaw_belt total_accel_belt
## 1 no 11 1.41 8.07 -94.4 3
## 2 no 11 1.41 8.07 -94.4 3
## 5 no 12 1.48 8.07 -94.4 3
## 9 no 12 1.43 8.16 -94.4 3
## 10 no 12 1.45 8.17 -94.4 3
## 14 no 12 1.42 8.21 -94.4 3
## kurtosis_roll_belt kurtosis_picth_belt kurtosis_yaw_belt skewness_roll_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## skewness_roll_belt.1 skewness_yaw_belt max_roll_belt max_picth_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## max_yaw_belt min_roll_belt min_pitch_belt min_yaw_belt amplitude_roll_belt
## 1 NA NA NA NA NA
## 2 NA NA NA NA NA
## 5 NA NA NA NA NA
## 9 NA NA NA NA NA
## 10 NA NA NA NA NA
## 14 NA NA NA NA NA
## amplitude_pitch_belt amplitude_yaw_belt var_total_accel_belt avg_roll_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## stddev_roll_belt var_roll_belt avg_pitch_belt stddev_pitch_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## var_pitch_belt avg_yaw_belt stddev_yaw_belt var_yaw_belt gyros_belt_x
## 1 NA NA NA NA 0.00
## 2 NA NA NA NA 0.02
## 5 NA NA NA NA 0.02
## 9 NA NA NA NA 0.02
## 10 NA NA NA NA 0.03
## 14 NA NA NA NA 0.02
## gyros_belt_y gyros_belt_z accel_belt_x accel_belt_y accel_belt_z
## 1 0.00 -0.02 -21 4 22
## 2 0.00 -0.02 -22 4 22
## 5 0.02 -0.02 -21 2 24
## 9 0.00 -0.02 -20 2 24
## 10 0.00 0.00 -21 4 22
## 14 0.00 -0.02 -22 4 21
## magnet_belt_x magnet_belt_y magnet_belt_z roll_arm pitch_arm yaw_arm
## 1 -3 599 -313 -128 22.5 -161
## 2 -7 608 -311 -128 22.5 -161
## 5 -6 600 -302 -128 22.1 -161
## 9 1 602 -312 -128 21.7 -161
## 10 -3 609 -308 -128 21.6 -161
## 14 -8 598 -310 -128 21.4 -161
## total_accel_arm var_accel_arm avg_roll_arm stddev_roll_arm var_roll_arm
## 1 34 NA NA NA NA
## 2 34 NA NA NA NA
## 5 34 NA NA NA NA
## 9 34 NA NA NA NA
## 10 34 NA NA NA NA
## 14 34 NA NA NA NA
## avg_pitch_arm stddev_pitch_arm var_pitch_arm avg_yaw_arm stddev_yaw_arm
## 1 NA NA NA NA NA
## 2 NA NA NA NA NA
## 5 NA NA NA NA NA
## 9 NA NA NA NA NA
## 10 NA NA NA NA NA
## 14 NA NA NA NA NA
## var_yaw_arm gyros_arm_x gyros_arm_y gyros_arm_z accel_arm_x accel_arm_y
## 1 NA 0.00 0.00 -0.02 -288 109
## 2 NA 0.02 -0.02 -0.02 -290 110
## 5 NA 0.00 -0.03 0.00 -289 111
## 9 NA 0.02 -0.03 -0.02 -288 109
## 10 NA 0.02 -0.03 -0.02 -288 110
## 14 NA 0.02 0.00 -0.03 -288 111
## accel_arm_z magnet_arm_x magnet_arm_y magnet_arm_z kurtosis_roll_arm
## 1 -123 -368 337 516 NA
## 2 -125 -369 337 513 NA
## 5 -123 -374 337 506 NA
## 9 -122 -369 341 518 NA
## 10 -124 -376 334 516 NA
## 14 -124 -371 331 523 NA
## kurtosis_picth_arm kurtosis_yaw_arm skewness_roll_arm skewness_pitch_arm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## skewness_yaw_arm max_roll_arm max_picth_arm max_yaw_arm min_roll_arm
## 1 NA NA NA NA NA
## 2 NA NA NA NA NA
## 5 NA NA NA NA NA
## 9 NA NA NA NA NA
## 10 NA NA NA NA NA
## 14 NA NA NA NA NA
## min_pitch_arm min_yaw_arm amplitude_roll_arm amplitude_pitch_arm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## amplitude_yaw_arm roll_dumbbell pitch_dumbbell yaw_dumbbell
## 1 NA 13.05217 -70.49400 -84.87394
## 2 NA 13.13074 -70.63751 -84.71065
## 5 NA 13.37872 -70.42856 -84.85306
## 9 NA 13.15463 -70.42520 -84.91563
## 10 NA 13.33034 -70.85059 -84.44602
## 14 NA 13.41048 -70.99594 -84.28005
## kurtosis_roll_dumbbell kurtosis_picth_dumbbell kurtosis_yaw_dumbbell
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## skewness_roll_dumbbell skewness_pitch_dumbbell skewness_yaw_dumbbell
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## max_roll_dumbbell max_picth_dumbbell max_yaw_dumbbell min_roll_dumbbell
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## min_pitch_dumbbell min_yaw_dumbbell amplitude_roll_dumbbell
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## amplitude_pitch_dumbbell amplitude_yaw_dumbbell total_accel_dumbbell
## 1 NA NA 37
## 2 NA NA 37
## 5 NA NA 37
## 9 NA NA 37
## 10 NA NA 37
## 14 NA NA 37
## var_accel_dumbbell avg_roll_dumbbell stddev_roll_dumbbell var_roll_dumbbell
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## avg_pitch_dumbbell stddev_pitch_dumbbell var_pitch_dumbbell avg_yaw_dumbbell
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## stddev_yaw_dumbbell var_yaw_dumbbell gyros_dumbbell_x gyros_dumbbell_y
## 1 NA NA 0.00 -0.02
## 2 NA NA 0.00 -0.02
## 5 NA NA 0.00 -0.02
## 9 NA NA 0.00 -0.02
## 10 NA NA 0.00 -0.02
## 14 NA NA 0.02 -0.02
## gyros_dumbbell_z accel_dumbbell_x accel_dumbbell_y accel_dumbbell_z
## 1 0.00 -234 47 -271
## 2 0.00 -233 47 -269
## 5 0.00 -233 48 -270
## 9 0.00 -232 47 -269
## 10 0.00 -235 48 -270
## 14 -0.02 -234 48 -268
## magnet_dumbbell_x magnet_dumbbell_y magnet_dumbbell_z roll_forearm
## 1 -559 293 -65 28.4
## 2 -555 296 -64 28.3
## 5 -554 292 -68 28.0
## 9 -549 292 -65 27.7
## 10 -558 291 -69 27.7
## 14 -554 295 -68 27.2
## pitch_forearm yaw_forearm kurtosis_roll_forearm kurtosis_picth_forearm
## 1 -63.9 -153 NA NA
## 2 -63.9 -153 NA NA
## 5 -63.9 -152 NA NA
## 9 -63.8 -152 NA NA
## 10 -63.8 -152 NA NA
## 14 -63.9 -151 NA NA
## kurtosis_yaw_forearm skewness_roll_forearm skewness_pitch_forearm
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## skewness_yaw_forearm max_roll_forearm max_picth_forearm max_yaw_forearm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## min_roll_forearm min_pitch_forearm min_yaw_forearm amplitude_roll_forearm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## amplitude_pitch_forearm amplitude_yaw_forearm total_accel_forearm
## 1 NA NA 36
## 2 NA NA 36
## 5 NA NA 36
## 9 NA NA 36
## 10 NA NA 36
## 14 NA NA 36
## var_accel_forearm avg_roll_forearm stddev_roll_forearm var_roll_forearm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## avg_pitch_forearm stddev_pitch_forearm var_pitch_forearm avg_yaw_forearm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## stddev_yaw_forearm var_yaw_forearm gyros_forearm_x gyros_forearm_y
## 1 NA NA 0.03 0.00
## 2 NA NA 0.02 0.00
## 5 NA NA 0.02 0.00
## 9 NA NA 0.03 0.00
## 10 NA NA 0.02 0.00
## 14 NA NA 0.00 -0.02
## gyros_forearm_z accel_forearm_x accel_forearm_y accel_forearm_z
## 1 -0.02 192 203 -215
## 2 -0.02 192 203 -216
## 5 -0.02 189 206 -214
## 9 -0.02 193 204 -214
## 10 -0.02 190 205 -215
## 14 -0.03 193 202 -214
## magnet_forearm_x magnet_forearm_y magnet_forearm_z classe
## 1 -17 654 476 A
## 2 -18 661 473 A
## 5 -17 655 473 A
## 9 -16 653 476 A
## 10 -22 656 473 A
## 14 -14 659 478 ACleaning Data:
Removing zero covariates, from this we will get to know about those variable who has very little variability and not likely good predictors.
nzv <- nearZeroVar(train1, saveMetrics = TRUE)
nzv## freqRatio percentUnique zeroVar nzv
## X 1.000000 1.000000e+02 FALSE FALSE
## user_name 1.102186 5.095109e-02 FALSE FALSE
## raw_timestamp_part_1 1.038462 7.107677e+00 FALSE FALSE
## raw_timestamp_part_2 1.250000 9.084579e+01 FALSE FALSE
## cvtd_timestamp 1.011111 1.698370e-01 FALSE FALSE
## new_window 48.898305 1.698370e-02 FALSE TRUE
## num_window 1.040000 7.269022e+00 FALSE FALSE
## roll_belt 1.097378 8.721128e+00 FALSE FALSE
## pitch_belt 1.061404 1.373981e+01 FALSE FALSE
## yaw_belt 1.101639 1.437670e+01 FALSE FALSE
## total_accel_belt 1.089256 2.462636e-01 FALSE FALSE
## kurtosis_roll_belt 2.000000 1.944633e+00 FALSE FALSE
## kurtosis_picth_belt 1.000000 1.672894e+00 FALSE FALSE
## kurtosis_yaw_belt 0.000000 0.000000e+00 TRUE TRUE
## skewness_roll_belt 1.000000 1.936141e+00 FALSE FALSE
## skewness_roll_belt.1 1.500000 1.757812e+00 FALSE FALSE
## skewness_yaw_belt 0.000000 0.000000e+00 TRUE TRUE
## max_roll_belt 1.666667 1.180367e+00 FALSE FALSE
## max_picth_belt 1.657895 1.698370e-01 FALSE FALSE
## max_yaw_belt 1.055556 4.840353e-01 FALSE FALSE
## min_roll_belt 1.333333 1.078465e+00 FALSE FALSE
## min_pitch_belt 2.218750 1.103940e-01 FALSE FALSE
## min_yaw_belt 1.055556 4.840353e-01 FALSE FALSE
## amplitude_roll_belt 1.100000 7.982337e-01 FALSE FALSE
## amplitude_pitch_belt 3.567568 9.341033e-02 FALSE FALSE
## amplitude_yaw_belt 0.000000 8.491848e-03 TRUE TRUE
## var_total_accel_belt 1.294118 3.991168e-01 FALSE FALSE
## avg_roll_belt 1.000000 1.129416e+00 FALSE FALSE
## stddev_roll_belt 1.103448 4.330842e-01 FALSE FALSE
## var_roll_belt 1.460000 5.264946e-01 FALSE FALSE
## avg_pitch_belt 1.200000 1.307745e+00 FALSE FALSE
## stddev_pitch_belt 1.026316 2.972147e-01 FALSE FALSE
## var_pitch_belt 1.363636 3.821332e-01 FALSE FALSE
## avg_yaw_belt 1.000000 1.384171e+00 FALSE FALSE
## stddev_yaw_belt 1.551724 3.651495e-01 FALSE FALSE
## var_yaw_belt 2.000000 8.237092e-01 FALSE FALSE
## gyros_belt_x 1.059627 1.095448e+00 FALSE FALSE
## gyros_belt_y 1.180442 5.604620e-01 FALSE FALSE
## gyros_belt_z 1.102589 1.341712e+00 FALSE FALSE
## accel_belt_x 1.035011 1.324728e+00 FALSE FALSE
## accel_belt_y 1.073876 1.163383e+00 FALSE FALSE
## accel_belt_z 1.085437 2.411685e+00 FALSE FALSE
## magnet_belt_x 1.085586 2.522079e+00 FALSE FALSE
## magnet_belt_y 1.103723 2.411685e+00 FALSE FALSE
## magnet_belt_z 1.000000 3.507133e+00 FALSE FALSE
## roll_arm 50.950000 1.944633e+01 FALSE FALSE
## pitch_arm 92.681818 2.253736e+01 FALSE FALSE
## yaw_arm 31.353846 2.145890e+01 FALSE FALSE
## total_accel_arm 1.005484 5.519701e-01 FALSE FALSE
## var_accel_arm 2.500000 1.961617e+00 FALSE FALSE
## avg_roll_arm 44.000000 1.638927e+00 FALSE TRUE
## stddev_roll_arm 44.000000 1.638927e+00 FALSE TRUE
## var_roll_arm 44.000000 1.638927e+00 FALSE TRUE
## avg_pitch_arm 44.000000 1.638927e+00 FALSE TRUE
## stddev_pitch_arm 44.000000 1.638927e+00 FALSE TRUE
## var_pitch_arm 44.000000 1.638927e+00 FALSE TRUE
## avg_yaw_arm 44.000000 1.638927e+00 FALSE TRUE
## stddev_yaw_arm 45.000000 1.630435e+00 FALSE TRUE
## var_yaw_arm 45.000000 1.630435e+00 FALSE TRUE
## gyros_arm_x 1.003106 5.239470e+00 FALSE FALSE
## gyros_arm_y 1.396226 3.048573e+00 FALSE FALSE
## gyros_arm_z 1.123810 1.944633e+00 FALSE FALSE
## accel_arm_x 1.099010 6.462296e+00 FALSE FALSE
## accel_arm_y 1.107692 4.449728e+00 FALSE FALSE
## accel_arm_z 1.157895 6.428329e+00 FALSE FALSE
## magnet_arm_x 1.000000 1.108186e+01 FALSE FALSE
## magnet_arm_y 1.076923 7.192595e+00 FALSE FALSE
## magnet_arm_z 1.061538 1.052989e+01 FALSE FALSE
## kurtosis_roll_arm 1.000000 1.630435e+00 FALSE FALSE
## kurtosis_picth_arm 1.000000 1.621943e+00 FALSE FALSE
## kurtosis_yaw_arm 2.000000 1.953125e+00 FALSE FALSE
## skewness_roll_arm 1.000000 1.630435e+00 FALSE FALSE
## skewness_pitch_arm 1.000000 1.621943e+00 FALSE FALSE
## skewness_yaw_arm 1.000000 1.944633e+00 FALSE FALSE
## max_roll_arm 14.666667 1.537024e+00 FALSE FALSE
## max_picth_arm 7.333333 1.409647e+00 FALSE FALSE
## max_yaw_arm 1.153846 4.161005e-01 FALSE FALSE
## min_roll_arm 14.666667 1.452106e+00 FALSE FALSE
## min_pitch_arm 14.666667 1.486073e+00 FALSE FALSE
## min_yaw_arm 1.200000 3.141984e-01 FALSE FALSE
## amplitude_roll_arm 22.000000 1.554008e+00 FALSE TRUE
## amplitude_pitch_arm 15.000000 1.520041e+00 FALSE FALSE
## amplitude_yaw_arm 1.400000 4.076087e-01 FALSE FALSE
## roll_dumbbell 1.064935 8.738961e+01 FALSE FALSE
## pitch_dumbbell 2.231707 8.519022e+01 FALSE FALSE
## yaw_dumbbell 1.093333 8.673573e+01 FALSE FALSE
## kurtosis_roll_dumbbell 2.000000 1.978601e+00 FALSE FALSE
## kurtosis_picth_dumbbell 2.000000 1.987092e+00 FALSE FALSE
## kurtosis_yaw_dumbbell 0.000000 0.000000e+00 TRUE TRUE
## skewness_roll_dumbbell 2.000000 1.978601e+00 FALSE FALSE
## skewness_pitch_dumbbell 1.000000 1.978601e+00 FALSE FALSE
## skewness_yaw_dumbbell 0.000000 0.000000e+00 TRUE TRUE
## max_roll_dumbbell 1.333333 1.825747e+00 FALSE FALSE
## max_picth_dumbbell 1.000000 1.817255e+00 FALSE FALSE
## max_yaw_dumbbell 1.083333 4.840353e-01 FALSE FALSE
## min_roll_dumbbell 1.333333 1.757812e+00 FALSE FALSE
## min_pitch_dumbbell 1.000000 1.851223e+00 FALSE FALSE
## min_yaw_dumbbell 1.083333 4.840353e-01 FALSE FALSE
## amplitude_roll_dumbbell 4.500000 1.919158e+00 FALSE FALSE
## amplitude_pitch_dumbbell 4.500000 1.910666e+00 FALSE FALSE
## amplitude_yaw_dumbbell 0.000000 8.491848e-03 TRUE TRUE
## total_accel_dumbbell 1.107807 3.566576e-01 FALSE FALSE
## var_accel_dumbbell 3.333333 1.910666e+00 FALSE FALSE
## avg_roll_dumbbell 1.500000 1.978601e+00 FALSE FALSE
## stddev_roll_dumbbell 9.000000 1.936141e+00 FALSE FALSE
## var_roll_dumbbell 9.000000 1.936141e+00 FALSE FALSE
## avg_pitch_dumbbell 1.500000 1.978601e+00 FALSE FALSE
## stddev_pitch_dumbbell 9.000000 1.936141e+00 FALSE FALSE
## var_pitch_dumbbell 9.000000 1.936141e+00 FALSE FALSE
## avg_yaw_dumbbell 1.500000 1.978601e+00 FALSE FALSE
## stddev_yaw_dumbbell 9.000000 1.936141e+00 FALSE FALSE
## var_yaw_dumbbell 9.000000 1.936141e+00 FALSE FALSE
## gyros_dumbbell_x 1.084034 1.919158e+00 FALSE FALSE
## gyros_dumbbell_y 1.314706 2.241848e+00 FALSE FALSE
## gyros_dumbbell_z 1.146974 1.630435e+00 FALSE FALSE
## accel_dumbbell_x 1.045685 3.422215e+00 FALSE FALSE
## accel_dumbbell_y 1.066225 3.744905e+00 FALSE FALSE
## accel_dumbbell_z 1.155405 3.345788e+00 FALSE FALSE
## magnet_dumbbell_x 1.160000 8.797554e+00 FALSE FALSE
## magnet_dumbbell_y 1.196429 6.929348e+00 FALSE FALSE
## magnet_dumbbell_z 1.060870 5.528193e+00 FALSE FALSE
## roll_forearm 11.658291 1.475883e+01 FALSE FALSE
## pitch_forearm 59.487179 2.107677e+01 FALSE FALSE
## yaw_forearm 16.330986 1.422385e+01 FALSE FALSE
## kurtosis_roll_forearm 2.000000 1.528533e+00 FALSE FALSE
## kurtosis_picth_forearm 1.000000 1.528533e+00 FALSE FALSE
## kurtosis_yaw_forearm 0.000000 0.000000e+00 TRUE TRUE
## skewness_roll_forearm 1.000000 1.537024e+00 FALSE FALSE
## skewness_pitch_forearm 4.000000 1.503057e+00 FALSE FALSE
## skewness_yaw_forearm 0.000000 0.000000e+00 TRUE TRUE
## max_roll_forearm 18.333333 1.426630e+00 FALSE FALSE
## max_picth_forearm 4.230769 8.661685e-01 FALSE FALSE
## max_yaw_forearm 1.100000 2.632473e-01 FALSE FALSE
## min_roll_forearm 27.500000 1.435122e+00 FALSE TRUE
## min_pitch_forearm 3.666667 8.916440e-01 FALSE FALSE
## min_yaw_forearm 1.100000 2.632473e-01 FALSE FALSE
## amplitude_roll_forearm 18.333333 1.426630e+00 FALSE FALSE
## amplitude_pitch_forearm 3.294118 9.341033e-01 FALSE FALSE
## amplitude_yaw_forearm 0.000000 8.491848e-03 TRUE TRUE
## total_accel_forearm 1.117105 5.774457e-01 FALSE FALSE
## var_accel_forearm 4.000000 1.978601e+00 FALSE FALSE
## avg_roll_forearm 27.500000 1.537024e+00 FALSE TRUE
## stddev_roll_forearm 57.000000 1.528533e+00 FALSE TRUE
## var_roll_forearm 57.000000 1.528533e+00 FALSE TRUE
## avg_pitch_forearm 55.000000 1.545516e+00 FALSE TRUE
## stddev_pitch_forearm 27.500000 1.537024e+00 FALSE TRUE
## var_pitch_forearm 55.000000 1.545516e+00 FALSE TRUE
## avg_yaw_forearm 55.000000 1.545516e+00 FALSE TRUE
## stddev_yaw_forearm 56.000000 1.537024e+00 FALSE TRUE
## var_yaw_forearm 56.000000 1.537024e+00 FALSE TRUE
## gyros_forearm_x 1.024169 2.377717e+00 FALSE FALSE
## gyros_forearm_y 1.107143 6.003736e+00 FALSE FALSE
## gyros_forearm_z 1.128814 2.386209e+00 FALSE FALSE
## accel_forearm_x 1.076923 6.547215e+00 FALSE FALSE
## accel_forearm_y 1.178571 8.160666e+00 FALSE FALSE
## accel_forearm_z 1.021277 4.662024e+00 FALSE FALSE
## magnet_forearm_x 1.040816 1.206692e+01 FALSE FALSE
## magnet_forearm_y 1.102041 1.531080e+01 FALSE FALSE
## magnet_forearm_z 1.025641 1.334918e+01 FALSE FALSE
## classe 1.469065 4.245924e-02 FALSE FALSERemoving the predictors which are not good:
train1<- train1[,!nzv$nzv]; valid<- valid[,!nzv$nzv]; test <- test[,!nzv$nzv];
dim(train1); dim(valid)## [1] 11776 130## [1] 7846 130head(train1)## X user_name raw_timestamp_part_1 raw_timestamp_part_2 cvtd_timestamp
## 1 1 carlitos 1323084231 788290 05/12/2011 11:23
## 2 2 carlitos 1323084231 808298 05/12/2011 11:23
## 5 5 carlitos 1323084232 196328 05/12/2011 11:23
## 9 9 carlitos 1323084232 484323 05/12/2011 11:23
## 10 10 carlitos 1323084232 484434 05/12/2011 11:23
## 14 14 carlitos 1323084232 576390 05/12/2011 11:23
## num_window roll_belt pitch_belt yaw_belt total_accel_belt kurtosis_roll_belt
## 1 11 1.41 8.07 -94.4 3 NA
## 2 11 1.41 8.07 -94.4 3 NA
## 5 12 1.48 8.07 -94.4 3 NA
## 9 12 1.43 8.16 -94.4 3 NA
## 10 12 1.45 8.17 -94.4 3 NA
## 14 12 1.42 8.21 -94.4 3 NA
## kurtosis_picth_belt skewness_roll_belt skewness_roll_belt.1 max_roll_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## max_picth_belt max_yaw_belt min_roll_belt min_pitch_belt min_yaw_belt
## 1 NA NA NA NA NA
## 2 NA NA NA NA NA
## 5 NA NA NA NA NA
## 9 NA NA NA NA NA
## 10 NA NA NA NA NA
## 14 NA NA NA NA NA
## amplitude_roll_belt amplitude_pitch_belt var_total_accel_belt avg_roll_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## stddev_roll_belt var_roll_belt avg_pitch_belt stddev_pitch_belt
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## var_pitch_belt avg_yaw_belt stddev_yaw_belt var_yaw_belt gyros_belt_x
## 1 NA NA NA NA 0.00
## 2 NA NA NA NA 0.02
## 5 NA NA NA NA 0.02
## 9 NA NA NA NA 0.02
## 10 NA NA NA NA 0.03
## 14 NA NA NA NA 0.02
## gyros_belt_y gyros_belt_z accel_belt_x accel_belt_y accel_belt_z
## 1 0.00 -0.02 -21 4 22
## 2 0.00 -0.02 -22 4 22
## 5 0.02 -0.02 -21 2 24
## 9 0.00 -0.02 -20 2 24
## 10 0.00 0.00 -21 4 22
## 14 0.00 -0.02 -22 4 21
## magnet_belt_x magnet_belt_y magnet_belt_z roll_arm pitch_arm yaw_arm
## 1 -3 599 -313 -128 22.5 -161
## 2 -7 608 -311 -128 22.5 -161
## 5 -6 600 -302 -128 22.1 -161
## 9 1 602 -312 -128 21.7 -161
## 10 -3 609 -308 -128 21.6 -161
## 14 -8 598 -310 -128 21.4 -161
## total_accel_arm var_accel_arm gyros_arm_x gyros_arm_y gyros_arm_z
## 1 34 NA 0.00 0.00 -0.02
## 2 34 NA 0.02 -0.02 -0.02
## 5 34 NA 0.00 -0.03 0.00
## 9 34 NA 0.02 -0.03 -0.02
## 10 34 NA 0.02 -0.03 -0.02
## 14 34 NA 0.02 0.00 -0.03
## accel_arm_x accel_arm_y accel_arm_z magnet_arm_x magnet_arm_y magnet_arm_z
## 1 -288 109 -123 -368 337 516
## 2 -290 110 -125 -369 337 513
## 5 -289 111 -123 -374 337 506
## 9 -288 109 -122 -369 341 518
## 10 -288 110 -124 -376 334 516
## 14 -288 111 -124 -371 331 523
## kurtosis_roll_arm kurtosis_picth_arm kurtosis_yaw_arm skewness_roll_arm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## skewness_pitch_arm skewness_yaw_arm max_roll_arm max_picth_arm max_yaw_arm
## 1 NA NA NA NA NA
## 2 NA NA NA NA NA
## 5 NA NA NA NA NA
## 9 NA NA NA NA NA
## 10 NA NA NA NA NA
## 14 NA NA NA NA NA
## min_roll_arm min_pitch_arm min_yaw_arm amplitude_pitch_arm amplitude_yaw_arm
## 1 NA NA NA NA NA
## 2 NA NA NA NA NA
## 5 NA NA NA NA NA
## 9 NA NA NA NA NA
## 10 NA NA NA NA NA
## 14 NA NA NA NA NA
## roll_dumbbell pitch_dumbbell yaw_dumbbell kurtosis_roll_dumbbell
## 1 13.05217 -70.49400 -84.87394 NA
## 2 13.13074 -70.63751 -84.71065 NA
## 5 13.37872 -70.42856 -84.85306 NA
## 9 13.15463 -70.42520 -84.91563 NA
## 10 13.33034 -70.85059 -84.44602 NA
## 14 13.41048 -70.99594 -84.28005 NA
## kurtosis_picth_dumbbell skewness_roll_dumbbell skewness_pitch_dumbbell
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## max_roll_dumbbell max_picth_dumbbell max_yaw_dumbbell min_roll_dumbbell
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## min_pitch_dumbbell min_yaw_dumbbell amplitude_roll_dumbbell
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## amplitude_pitch_dumbbell total_accel_dumbbell var_accel_dumbbell
## 1 NA 37 NA
## 2 NA 37 NA
## 5 NA 37 NA
## 9 NA 37 NA
## 10 NA 37 NA
## 14 NA 37 NA
## avg_roll_dumbbell stddev_roll_dumbbell var_roll_dumbbell avg_pitch_dumbbell
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## stddev_pitch_dumbbell var_pitch_dumbbell avg_yaw_dumbbell
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## stddev_yaw_dumbbell var_yaw_dumbbell gyros_dumbbell_x gyros_dumbbell_y
## 1 NA NA 0.00 -0.02
## 2 NA NA 0.00 -0.02
## 5 NA NA 0.00 -0.02
## 9 NA NA 0.00 -0.02
## 10 NA NA 0.00 -0.02
## 14 NA NA 0.02 -0.02
## gyros_dumbbell_z accel_dumbbell_x accel_dumbbell_y accel_dumbbell_z
## 1 0.00 -234 47 -271
## 2 0.00 -233 47 -269
## 5 0.00 -233 48 -270
## 9 0.00 -232 47 -269
## 10 0.00 -235 48 -270
## 14 -0.02 -234 48 -268
## magnet_dumbbell_x magnet_dumbbell_y magnet_dumbbell_z roll_forearm
## 1 -559 293 -65 28.4
## 2 -555 296 -64 28.3
## 5 -554 292 -68 28.0
## 9 -549 292 -65 27.7
## 10 -558 291 -69 27.7
## 14 -554 295 -68 27.2
## pitch_forearm yaw_forearm kurtosis_roll_forearm kurtosis_picth_forearm
## 1 -63.9 -153 NA NA
## 2 -63.9 -153 NA NA
## 5 -63.9 -152 NA NA
## 9 -63.8 -152 NA NA
## 10 -63.8 -152 NA NA
## 14 -63.9 -151 NA NA
## skewness_roll_forearm skewness_pitch_forearm max_roll_forearm
## 1 NA NA NA
## 2 NA NA NA
## 5 NA NA NA
## 9 NA NA NA
## 10 NA NA NA
## 14 NA NA NA
## max_picth_forearm max_yaw_forearm min_pitch_forearm min_yaw_forearm
## 1 NA NA NA NA
## 2 NA NA NA NA
## 5 NA NA NA NA
## 9 NA NA NA NA
## 10 NA NA NA NA
## 14 NA NA NA NA
## amplitude_roll_forearm amplitude_pitch_forearm total_accel_forearm
## 1 NA NA 36
## 2 NA NA 36
## 5 NA NA 36
## 9 NA NA 36
## 10 NA NA 36
## 14 NA NA 36
## var_accel_forearm gyros_forearm_x gyros_forearm_y gyros_forearm_z
## 1 NA 0.03 0.00 -0.02
## 2 NA 0.02 0.00 -0.02
## 5 NA 0.02 0.00 -0.02
## 9 NA 0.03 0.00 -0.02
## 10 NA 0.02 0.00 -0.02
## 14 NA 0.00 -0.02 -0.03
## accel_forearm_x accel_forearm_y accel_forearm_z magnet_forearm_x
## 1 192 203 -215 -17
## 2 192 203 -216 -18
## 5 189 206 -214 -17
## 9 193 204 -214 -16
## 10 190 205 -215 -22
## 14 193 202 -214 -14
## magnet_forearm_y magnet_forearm_z classe
## 1 654 476 A
## 2 661 473 A
## 5 655 473 A
## 9 653 476 A
## 10 656 473 A
## 14 659 478 ARemoving NAs
cleantr <- train1[, colSums(is.na(train1)) == 0]
cleanVa <- valid[, colSums(is.na(valid)) == 0]
test<- test[,colSums(is.na(valid)) == 0]Removing the class column from Training Set.
finalTrain <- cleantr[, -(1:5)]
finalValid <- cleanVa[,-(1:5)]Creating model:
Decision Tree
set.seed(12345)
modFit <- train(classe ~ .,method="rpart", data = finalTrain)
modFit## CART
##
## 11776 samples
## 53 predictor
## 5 classes: 'A', 'B', 'C', 'D', 'E'
##
## No pre-processing
## Resampling: Bootstrapped (25 reps)
## Summary of sample sizes: 11776, 11776, 11776, 11776, 11776, 11776, ...
## Resampling results across tuning parameters:
##
## cp Accuracy Kappa
## 0.02432368 0.6310912 0.53332175
## 0.04358488 0.5114667 0.36147484
## 0.11449929 0.3270387 0.06354879
##
## Accuracy was used to select the optimal model using the largest value.
## The final value used for the model was cp = 0.02432368.Creating Decision Tree:
fancyRpartPlot(modFit$finalModel)
Now, Predicting class using ValidationSet:
pred <- predict(modFit, newdata = finalValid)
cnfMatrix <- confusionMatrix(pred, data = as.factor(finalValid$classe))
cnfMatrix## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1886 152 189 0 5
## B 286 824 408 0 0
## C 165 123 1080 0 0
## D 298 327 600 0 61
## E 90 249 215 0 888
##
## Overall Statistics
##
## Accuracy : 0.5962
## 95% CI : (0.5853, 0.6071)
## No Information Rate : 0.3473
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.4838
##
## Mcnemar's Test P-Value : < 2.2e-16
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 0.6921 0.4919 0.4334 NA 0.9308
## Specificity 0.9324 0.8875 0.9462 0.8361 0.9196
## Pos Pred Value 0.8450 0.5428 0.7895 NA 0.6158
## Neg Pred Value 0.8506 0.8655 0.7820 NA 0.9897
## Prevalence 0.3473 0.2135 0.3176 0.0000 0.1216
## Detection Rate 0.2404 0.1050 0.1376 0.0000 0.1132
## Detection Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Balanced Accuracy 0.8123 0.6897 0.6898 NA 0.9252plot(cnfMatrix$table, col = cnfMatrix$byClass, main = paste("Decision Tree - Accuracy =", round(cnfMatrix$overall['Accuracy'], 3)))
Random Forest
control <- trainControl(method="cv", number=5, verboseIter=TRUE)
modFit1 <- train(classe ~ .,method="rf", trControl=control, data = finalTrain)## + Fold1: mtry= 2
## - Fold1: mtry= 2
## + Fold1: mtry=27
## - Fold1: mtry=27
## + Fold1: mtry=53
## - Fold1: mtry=53
## + Fold2: mtry= 2
## - Fold2: mtry= 2
## + Fold2: mtry=27
## - Fold2: mtry=27
## + Fold2: mtry=53
## - Fold2: mtry=53
## + Fold3: mtry= 2
## - Fold3: mtry= 2
## + Fold3: mtry=27
## - Fold3: mtry=27
## + Fold3: mtry=53
## - Fold3: mtry=53
## + Fold4: mtry= 2
## - Fold4: mtry= 2
## + Fold4: mtry=27
## - Fold4: mtry=27
## + Fold4: mtry=53
## - Fold4: mtry=53
## + Fold5: mtry= 2
## - Fold5: mtry= 2
## + Fold5: mtry=27
## - Fold5: mtry=27
## + Fold5: mtry=53
## - Fold5: mtry=53
## Aggregating results
## Selecting tuning parameters
## Fitting mtry = 27 on full training setmodFit1## Random Forest
##
## 11776 samples
## 53 predictor
## 5 classes: 'A', 'B', 'C', 'D', 'E'
##
## No pre-processing
## Resampling: Cross-Validated (5 fold)
## Summary of sample sizes: 9421, 9421, 9420, 9421, 9421
## Resampling results across tuning parameters:
##
## mtry Accuracy Kappa
## 2 0.9909985 0.9886128
## 27 0.9964334 0.9954886
## 53 0.9935461 0.9918360
##
## Accuracy was used to select the optimal model using the largest value.
## The final value used for the model was mtry = 27.Now, Predicting class using ValidationSet:
pred1 <- predict(modFit1, newdata = finalValid)
cnfMatrix1 <- confusionMatrix(pred1, data = as.factor(finalValid$classe))
cnfMatrix1## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 2232 0 0 0 0
## B 2 1516 0 0 0
## C 0 1 1367 0 0
## D 0 0 3 1283 0
## E 0 0 0 3 1439
##
## Overall Statistics
##
## Accuracy : 0.9989
## 95% CI : (0.9978, 0.9995)
## No Information Rate : 0.2847
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.9985
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 0.9991 0.9993 0.9978 0.9977 1.0000
## Specificity 1.0000 0.9997 0.9998 0.9995 0.9995
## Pos Pred Value 1.0000 0.9987 0.9993 0.9977 0.9979
## Neg Pred Value 0.9996 0.9998 0.9995 0.9995 1.0000
## Prevalence 0.2847 0.1933 0.1746 0.1639 0.1834
## Detection Rate 0.2845 0.1932 0.1742 0.1635 0.1834
## Detection Prevalence 0.2845 0.1935 0.1744 0.1639 0.1838
## Balanced Accuracy 0.9996 0.9995 0.9988 0.9986 0.9998plot(cnfMatrix1$table, col = cnfMatrix1$byClass, main = paste("Random Forest - Accuracy =", round(cnfMatrix1$overall['Accuracy'], 3)))
Implementing Model on a Test Set:
As the accuracy of model created with Random Forest is high as compare to model created with Decision Tree. So, we will use the Random Forest Model to predict the Test Set class.
predResult <- predict(modFit1, newdata = test)
predResult## [1] B A B A A E D B A A B C B A E E A B B B
## Levels: A B C D E