1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 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
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).
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
setInternet2(TRUE)
download.file(url = "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv", destfile = "pml-training.csv")
download.file(url = "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv", destfile = "pml-testing.csv")
pmltrain <- read.csv("pml-training.csv")
pmltest <- read.csv("pml-testing.csv")Create training, test and validation sets
library(caret)## Loading required package: lattice
## Loading required package: ggplot2library(RANN)
library(randomForest)## randomForest 4.6-10
## Type rfNews() to see new features/changes/bug fixes.library(rpart)
library(rpart.plot)
trainidx <- createDataPartition(pmltrain$classe,p=.9,list=FALSE)
traindata = pmltrain[trainidx,]
testdata = pmltrain[-trainidx,]
set.seed(32768)
nzv <- nearZeroVar(traindata)
trainnzv <- traindata[-nzv]
testnzv <- testdata[-nzv]
pmltestnzv <- pmltest[-nzv]
dim(trainnzv)## [1] 17662 102dim(testnzv)## [1] 1960 102dim(pmltestnzv)## [1] 20 102ftridx <- which(lapply(trainnzv,class) %in% c('numeric'))
trainnzv1 <- preProcess(trainnzv[,ftridx], method=c('knnImpute'))
ftridx## [1] 7 8 9 11 13 15 17 18 19 20 21 22 23 24 25 26 27
## [18] 28 29 36 37 38 40 41 42 43 50 52 53 56 57 58 59 60
## [35] 61 62 63 64 66 67 68 69 70 71 72 73 74 75 76 77 78
## [52] 84 85 86 87 88 89 90 92 93 94 95 100 101trainnzv1##
## Call:
## preProcess.default(x = trainnzv[, ftridx], method = c("knnImpute"))
##
## Created from 368 samples and 64 variables
## Pre-processing: 5 nearest neighbor imputation, scaled, centeredpred1 <- predict(trainnzv1, trainnzv[,ftridx])
predtrain <- cbind(trainnzv$classe,pred1)
names(predtrain)[1] <- 'classe'
predtrain[is.na(predtrain)] <- 0
pred2 <- predict(trainnzv1, testnzv[,ftridx])
predtest <- cbind(testnzv$classe, pred2)
names(predtest)[1] <- 'classe'
predtest[is.na(predtest)] <- 0
predpmltest <- predict(trainnzv1,pmltestnzv[,ftridx] )
dim(predtrain)## [1] 17662 65dim(predtest)## [1] 1960 65dim(predpmltest)## [1] 20 64model <- randomForest(classe~.,data=predtrain)
predtrain1 <- predict(model, predtrain)
print(table(predtrain1, predtrain$classe))##
## predtrain1 A B C D E
## A 5022 0 0 0 0
## B 0 3418 0 0 0
## C 0 0 3080 0 0
## D 0 0 0 2895 0
## E 0 0 0 0 3247training <- as.data.frame(table(predtrain1, predtrain$classe))
#qplot(training)
predtest1 <- predict(model, predtest)
print(table(predtest1, predtest$classe))##
## predtest1 A B C D E
## A 556 1 0 0 0
## B 2 378 5 0 0
## C 0 0 335 3 0
## D 0 0 2 318 2
## E 0 0 0 0 358str(predpmltest)## 'data.frame': 20 obs. of 64 variables:
## $ roll_belt : num 0.932 -1.012 -1.015 0.964 -1.007 ...
## $ pitch_belt : num 1.1932 0.205 0.0689 -1.8699 0.1363 ...
## $ yaw_belt : num 0.0648 -0.818 -0.8138 1.8141 -0.8149 ...
## $ max_roll_belt : num 0.0211 -0.8734 -0.8723 1.7935 -0.8896 ...
## $ min_roll_belt : num 0.0482 -0.8428 -0.8422 1.8326 -0.8602 ...
## $ amplitude_roll_belt : num -0.0944 -0.1392 -0.1377 -0.0759 -0.1354 ...
## $ var_total_accel_belt : num -0.321 -0.338 -0.347 -0.382 -0.38 ...
## $ avg_roll_belt : num 0.848 -1.085 -1.09 0.886 -1.076 ...
## $ stddev_roll_belt : num -0.413 -0.477 -0.437 -0.421 -0.445 ...
## $ var_roll_belt : num -0.329 -0.333 -0.332 -0.327 -0.332 ...
## $ avg_pitch_belt : num 1.113 0.199 0.192 -1.855 0.171 ...
## $ stddev_pitch_belt : num -0.5452 -0.0739 -0.1996 -0.4195 -0.1493 ...
## $ var_pitch_belt : num -0.401 -0.226 -0.284 -0.342 -0.259 ...
## $ avg_yaw_belt : num 0.0385 -0.8616 -0.8602 1.8249 -0.8777 ...
## $ stddev_yaw_belt : num -0.0902 -0.1235 -0.1216 -0.0643 -0.1196 ...
## $ var_yaw_belt : num -0.068 -0.0682 -0.0682 -0.0677 -0.0682 ...
## $ gyros_belt_x : num -2.38 -0.263 0.266 0.555 0.17 ...
## $ gyros_belt_y : num -0.766 -0.766 -0.255 0.895 -0.255 ...
## $ gyros_belt_z : num -1.358 0.248 0.66 -0.123 0.536 ...
## $ roll_arm : num 0.314 -0.246 -0.246 -1.744 0.801 ...
## $ pitch_arm : num -0.757 0.15 0.15 1.944 0.24 ...
## $ yaw_arm : num 2.50109 0.00913 0.00913 -1.97883 1.43711 ...
## $ var_accel_arm : num 0.6537 -0.9021 0.0811 -0.4641 0.2415 ...
## $ gyros_arm_x : num -0.8488 -0.6079 1.0332 0.0897 -1.0043 ...
## $ gyros_arm_y : num 0.865 1.3 -1.298 -0.299 1.229 ...
## $ gyros_arm_z : num -0.812 -1.263 1.55 1.171 -1.461 ...
## $ max_picth_arm : num 0.891 -0.544 -0.544 -1.318 0.839 ...
## $ min_roll_arm : num -0.471 0.796 0.796 0.935 -0.665 ...
## $ min_pitch_arm : num 0.674 0.59 0.59 -1.333 0.497 ...
## $ roll_dumbbell : num -0.591 0.44 0.478 0.278 -1.786 ...
## $ pitch_dumbbell : num 0.97 -1.155 -1.093 -0.516 -1.148 ...
## $ yaw_dumbbell : num 1.512 -0.935 -0.931 -1.272 -0.191 ...
## $ max_roll_dumbbell : num 0.695 -1.272 -1.107 -0.842 0.211 ...
## $ max_picth_dumbbell : num 1.15 -1.114 -0.993 -1.298 0.642 ...
## $ min_roll_dumbbell : num 1.499 -0.43 -0.593 0.1 0.616 ...
## $ min_pitch_dumbbell : num 1.844 -0.676 -0.684 -0.968 0.761 ...
## $ amplitude_roll_dumbbell : num -0.343 -0.832 -0.587 -0.795 -0.205 ...
## $ amplitude_pitch_dumbbell: num -0.433 -0.8329 -0.6503 -0.7677 0.0616 ...
## $ var_accel_dumbbell : num -0.396 -0.443 -0.3 -0.461 -0.363 ...
## $ avg_roll_dumbbell : num -0.4 0.494 0.593 0.4 -0.644 ...
## $ stddev_roll_dumbbell : num -0.345 -0.738 -0.522 -0.639 -0.196 ...
## $ var_roll_dumbbell : num -0.381 -0.447 -0.41 -0.437 -0.308 ...
## $ avg_pitch_dumbbell : num 1.284 -1.212 -1.149 -0.58 0.431 ...
## $ stddev_pitch_dumbbell : num -0.374 -0.82 -0.641 -0.793 -0.138 ...
## $ var_pitch_dumbbell : num -0.42 -0.511 -0.487 -0.511 -0.235 ...
## $ avg_yaw_dumbbell : num 1.588 -0.984 -0.928 -1.24 0.691 ...
## $ stddev_yaw_dumbbell : num -0.4634 -0.7937 -0.6342 -0.7294 0.0308 ...
## $ var_yaw_dumbbell : num -0.413 -0.465 -0.436 -0.459 -0.198 ...
## $ gyros_dumbbell_x : num 0.3029 0.1136 0.1451 -0.0379 0.082 ...
## $ gyros_dumbbell_y : num 0.02292 0.00681 0.15181 -0.10597 -0.83098 ...
## $ gyros_dumbbell_z : num -0.2005 -0.242 -0.0883 0.0737 -0.1382 ...
## $ magnet_dumbbell_z : num -0.7296 -0.5866 -0.036 0.0498 1.9019 ...
## $ roll_forearm : num 0.993 0.696 0.9 -0.314 -1.944 ...
## $ pitch_forearm : num 1.372 -1.001 -1.533 -0.377 -0.453 ...
## $ yaw_forearm : num 1.325 0.84 0.714 -0.187 -0.651 ...
## $ max_picth_forearm : num 0.647 0.551 0.538 -0.835 -0.113 ...
## $ min_pitch_forearm : num 0.823 1.353 0.896 0.545 -0.453 ...
## $ amplitude_pitch_forearm : num -0.19 -0.646 -0.315 -0.949 0.263 ...
## $ var_accel_forearm : num -0.648 -0.685 -0.767 -0.684 0.11 ...
## $ gyros_forearm_x : num 0.8953 1.4815 0.0314 1.8826 -1.4032 ...
## $ gyros_forearm_y : num -1.073 -0.898 -0.273 0.192 0.948 ...
## $ gyros_forearm_z : num -0.4052 -0.182 0.0683 0.8955 0.3513 ...
## $ magnet_forearm_y : num 0.0753 0.8064 0.6236 0.7907 -2.2948 ...
## $ magnet_forearm_z : num 0.604 1.298 1.054 0.343 -0.824 ...predanswers <- predict(model, predpmltest)
predanswers## 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
## 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 Epredanswers <- predict(model, predpmltest)
predanswers## 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
## 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 Epml_write_files = function(x){
n = length(x)
for(i in 1:n){
filename = paste0("problem_id_",i,".txt")
write.table(x[i],file=filename,quote=FALSE,row.names=FALSE,col.names=FALSE)
}
}
pml_write_files(as.character(predanswers))