The instructions for this project have been the following:
“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."
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://web.archive.org/web/20161224072740/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
The data for this project come from this source: http://web.archive.org/web/20161224072740/http:/groupware.les.inf.puc-rio.br/har.
library(knitr)
library(caret)
library(rpart)
library(rpart.plot)
library(randomForest)
library(corrplot)
library(rattle)Set the URL for the download
UrlTrain <- "http://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv"
UrlTest <- "http://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv"Download the datasets
training <- read.csv(url(UrlTrain))
testing <- read.csv(url(UrlTest))
dim(training)## [1] 19622 160dim(testing)## [1] 20 160Removing Variables which are having Nearly Zero Variance.
nzv <- nearZeroVar(training)
training_data <- training[,-nzv]
testing_data <- testing[,-nzv]
dim(training_data)## [1] 19622 100dim(testing_data)## [1] 20 100Removing NA Values of Variables.
allNA <- sapply(training_data, function(x) mean(is.na(x))) > 0.95
training_data <- training_data[,allNA == FALSE]
testing_data <- testing_data[,allNA == FALSE]
dim(training_data)## [1] 19622 59dim(testing_data)## [1] 20 59Remove identification only variables (columns 1 to 6)
training_data<- training_data[, 7:59]
testing_data<- testing_data[, 7:59]
dim(training_data)## [1] 19622 53dim(testing_data)## [1] 20 53I will use three model:
Data Partioning:
inTrain<- createDataPartition(training_data$classe, p=0.7, list=FALSE)
training<- training_data[inTrain,]
testing<- training_data[-inTrain,]
dim(training)## [1] 13737 53dim(testing)## [1] 5885 53set.seed(11111)
RF_model<- train(classe ~. , data=training, method= "rf", ntree=100)
RF_prediction<- predict(RF_model, testing)
RF_cm<-confusionMatrix(RF_prediction, testing$classe)
RF_cm## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1674 8 0 0 0
## B 0 1131 6 1 1
## C 0 0 1014 10 5
## D 0 0 6 953 1
## E 0 0 0 0 1075
##
## Overall Statistics
##
## Accuracy : 0.9935
## 95% CI : (0.9911, 0.9954)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.9918
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 1.0000 0.9930 0.9883 0.9886 0.9935
## Specificity 0.9981 0.9983 0.9969 0.9986 1.0000
## Pos Pred Value 0.9952 0.9930 0.9854 0.9927 1.0000
## Neg Pred Value 1.0000 0.9983 0.9975 0.9978 0.9985
## Prevalence 0.2845 0.1935 0.1743 0.1638 0.1839
## Detection Rate 0.2845 0.1922 0.1723 0.1619 0.1827
## Detection Prevalence 0.2858 0.1935 0.1749 0.1631 0.1827
## Balanced Accuracy 0.9991 0.9956 0.9926 0.9936 0.9968From the Random Forest Model we see the prediction accuracy is 99%.
#Fit the model and plot
set.seed(3333)
DT_model<- train(classe ~. , data=training, method= "rpart")
#Prediction
DT_prediction<- predict(DT_model, testing)
confusionMatrix(DT_prediction, testing$classe)## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1519 457 474 450 150
## B 34 404 31 163 153
## C 119 278 521 351 292
## D 0 0 0 0 0
## E 2 0 0 0 487
##
## Overall Statistics
##
## Accuracy : 0.498
## 95% CI : (0.4852, 0.5109)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.3441
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 0.9074 0.35470 0.50780 0.0000 0.45009
## Specificity 0.6364 0.91972 0.78596 1.0000 0.99958
## Pos Pred Value 0.4980 0.51465 0.33376 NaN 0.99591
## Neg Pred Value 0.9453 0.85588 0.88321 0.8362 0.88973
## Prevalence 0.2845 0.19354 0.17434 0.1638 0.18386
## Detection Rate 0.2581 0.06865 0.08853 0.0000 0.08275
## Detection Prevalence 0.5183 0.13339 0.26525 0.0000 0.08309
## Balanced Accuracy 0.7719 0.63721 0.64688 0.5000 0.72484From the Decision Tree Model we see the prediction accuracy is 50%
We conclude that, Random Forest is more accurate than Decision Tree Model.
prediction_test<- predict(RF_model, testing_data)
prediction_test## [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