Activity recognition - predictive analytics
Activity recognition - predictive analytics
Dhyanesh babu 6 April 2018
Summary
sing 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).
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://groupware.les.inf.puc-rio.br/har.
Steps performed
- Loading the data
- Cleaning the data
- Model selection
- Predicting test output
loading the data
library(caret)## Loading required package: lattice## Loading required package: ggplot2library(randomForest)## randomForest 4.6-14## Type rfNews() to see new features/changes/bug fixes.##
## Attaching package: 'randomForest'## The following object is masked from 'package:ggplot2':
##
## marginlibrary(plyr)
training_URL <- "http://d396qusza40orc.cloudfront.net/predmachlearn/pml-training.csv"
testing_URL <- "http://d396qusza40orc.cloudfront.net/predmachlearn/pml-testing.csv"
training <- read.csv(url(training_URL))
testing <- read.csv(url(testing_URL))Cleaning the data
set.seed(32233)
#Remove few columns
training <- subset(training, select=-c(1:7))
testing <- subset(testing, select=-c(1:7))
threshold_val <- 0.95 * dim(training)[1]
# Remove 95 % of the column values are NA
include_columns <- !apply(training, 2, function(y) sum(is.na(y)) > threshold_val || sum(y=="") > threshold_val)
training <- training[, include_columns]
dim(training)## [1] 19622 53#variance near zero columns
nearZvar <- nearZeroVar(training, saveMetrics = TRUE)
training <- training[ , nearZvar$nzv==FALSE]
dim(training)## [1] 19622 53spliting the data
#split the dataframe
label <- createDataPartition(training$classe, p = 0.7, list = FALSE)
train <- training[label, ]
test <- training[-label, ]
library(corrplot)## corrplot 0.84 loadedcorrMat <- cor(train[,-53])
corrplot(corrMat, method = "color", type = "lower", tl.cex = 0.8, tl.col = rgb(0,0,0))
Model selection
Decision Tree
Traning
library(rpart)
library(rpart.plot)
library(rattle)## Rattle: A free graphical interface for data science with R.
## Version 5.1.0 Copyright (c) 2006-2017 Togaware Pty Ltd.
## Type 'rattle()' to shake, rattle, and roll your data.##
## Attaching package: 'rattle'## The following object is masked from 'package:randomForest':
##
## importanceset.seed(13908)
model_decision_tree <- rpart(classe ~ ., data = train, method = "class")
fancyRpartPlot(model_decision_tree)## Warning: labs do not fit even at cex 0.15, there may be some overplotting
predict_decision_tree <- predict(model_decision_tree, test, type = "class")
conf_matric_decision_tree <- confusionMatrix(predict_decision_tree, test$classe)
conf_matric_decision_tree## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1500 183 21 59 10
## B 67 664 50 65 85
## C 39 196 875 104 151
## D 55 74 77 660 89
## E 13 22 3 76 747
##
## Overall Statistics
##
## Accuracy : 0.7555
## 95% CI : (0.7443, 0.7664)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.6904
## Mcnemar's Test P-Value : < 2.2e-16
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 0.8961 0.5830 0.8528 0.6846 0.6904
## Specificity 0.9352 0.9437 0.8992 0.9401 0.9763
## Pos Pred Value 0.8460 0.7132 0.6410 0.6911 0.8676
## Neg Pred Value 0.9577 0.9041 0.9666 0.9383 0.9333
## Prevalence 0.2845 0.1935 0.1743 0.1638 0.1839
## Detection Rate 0.2549 0.1128 0.1487 0.1121 0.1269
## Detection Prevalence 0.3013 0.1582 0.2319 0.1623 0.1463
## Balanced Accuracy 0.9156 0.7634 0.8760 0.8124 0.8333Random forest
library(caret)
set.seed(13908)
control <- trainControl(method = "cv", number = 3, verboseIter=FALSE)
model_Random_Forest <- train(classe ~ ., data = train, method = "rf", trControl = control)
model_Random_Forest$finalModel##
## Call:
## randomForest(x = x, y = y, mtry = param$mtry)
## Type of random forest: classification
## Number of trees: 500
## No. of variables tried at each split: 27
##
## OOB estimate of error rate: 0.73%
## Confusion matrix:
## A B C D E class.error
## A 3899 3 2 0 2 0.001792115
## B 18 2634 5 1 0 0.009029345
## C 0 16 2370 10 0 0.010851419
## D 0 1 28 2221 2 0.013765542
## E 0 1 4 7 2513 0.004752475predict
predict_random_forest <- predict(model_Random_Forest, test)
conf_matrix_Random_Forest <- confusionMatrix(predict_random_forest, test$classe)
conf_matrix_Random_Forest## Confusion Matrix and Statistics
##
## Reference
## Prediction A B C D E
## A 1674 11 0 0 0
## B 0 1126 5 1 1
## C 0 2 1017 6 3
## D 0 0 4 956 5
## E 0 0 0 1 1073
##
## Overall Statistics
##
## Accuracy : 0.9934
## 95% CI : (0.991, 0.9953)
## No Information Rate : 0.2845
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.9916
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: A Class: B Class: C Class: D Class: E
## Sensitivity 1.0000 0.9886 0.9912 0.9917 0.9917
## Specificity 0.9974 0.9985 0.9977 0.9982 0.9998
## Pos Pred Value 0.9935 0.9938 0.9893 0.9907 0.9991
## Neg Pred Value 1.0000 0.9973 0.9981 0.9984 0.9981
## Prevalence 0.2845 0.1935 0.1743 0.1638 0.1839
## Detection Rate 0.2845 0.1913 0.1728 0.1624 0.1823
## Detection Prevalence 0.2863 0.1925 0.1747 0.1640 0.1825
## Balanced Accuracy 0.9987 0.9936 0.9945 0.9949 0.9957predicting test output
predict_random_forest<- predict(model_Random_Forest, testing)
predict_random_forest## [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