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 (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 goal of the project is to predict the manner in which they did the exercise. This is the “classe” variable in the training set.
archivo = "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv"
file <- download.file(archivo, destfile = "pml.csv")
data <- read.csv("pml.csv", stringsAsFactors = F)
Factor the variable classe
str(data$classe)
## chr [1:19622] "A" "A" "A" "A" "A" "A" "A" "A" "A" "A" "A" "A" "A" "A" ...
unique(data$classe)
## [1] "A" "B" "C" "D" "E"
data$classe <- factor(data$classe)
Check the proportion of values of the variable class
round(prop.table(table(data$classe)), 2)
##
## A B C D E
## 0.28 0.19 0.17 0.16 0.18
At this moment we have 160 variables. Next, we will eliminate variables that are not significant for the analysis.
data_clean <- data
dim(data_clean)
## [1] 19622 160
data_clean <- data_clean[,-c(1:7)]
dim(data_clean)
## [1] 19622 153
nzv <- nearZeroVar(data_clean, saveMetrics = T)
data_clean <- data_clean[,nzv$nzv==FALSE]
dim(data_clean)
## [1] 19622 94
AllNA <- sapply(data_clean, function(x) mean(is.na(x))) > 0.95
data_clean <- data_clean[, AllNA==FALSE]
dim(data_clean)
## [1] 19622 53
We have left 19,000 observations of 53 variables. We divide the dataset into training and validation
set.seed(3455)
inTrain <- createDataPartition(y = data_clean$classe, p = 0.75, list = F)
training <- data_clean[inTrain,]
validation <- data_clean[-inTrain,]
# Comprobamos que se mantienen las proporciones
round(prop.table(table(training$classe)), 2)
##
## A B C D E
## 0.28 0.19 0.17 0.16 0.18
round(prop.table(table(validation$classe)), 2)
##
## A B C D E
## 0.28 0.19 0.17 0.16 0.18
We will use Random Forest beccause it is an algorithm that performs well on most problems, can handle categorical or continuous features, and can be used on very large datasets.
rf <- randomForest::randomForest(classe~., data = training)
prediction <- predict(rf, validation)
confusionMatrix(prediction, validation$classe)
## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1394 6 0 0 0
## B 1 943 3 0 0
## C 0 0 852 7 0
## D 0 0 0 797 0
## E 0 0 0 0 901
##
## Overall Statistics
##
## Accuracy : 0.9965
## 95% CI : (0.9945, 0.998)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.9956
##
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 0.9993 0.9937 0.9965 0.9913 1.0000
## Specificity 0.9983 0.9990 0.9983 1.0000 1.0000
## Pos Pred Value 0.9957 0.9958 0.9919 1.0000 1.0000
## Neg Pred Value 0.9997 0.9985 0.9993 0.9983 1.0000
## Prevalence 0.2845 0.1935 0.1743 0.1639 0.1837
## Detection Rate 0.2843 0.1923 0.1737 0.1625 0.1837
## Detection Prevalence 0.2855 0.1931 0.1752 0.1625 0.1837
## Balanced Accuracy 0.9988 0.9963 0.9974 0.9956 1.0000
archivo = "https://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv"
file <- download.file(archivo, destfile = "pml_testing.csv")
data_testing <- read.csv("pml_testing.csv", stringsAsFactors = F)
prediction <- predict(rf, data_testing)
prediction
## 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