About the App

This application is a simple demonstration of running ‘random forest’ classification. It shows the impact of a subset of choices made for training the model and their impact.

The application uses the ‘IRIS’ dataset which has the following structure:

str(iris)

## 'data.frame':    150 obs. of  5 variables:
##  $ Sepal.Length: num  5.1 4.9 4.7 4.6 5 5.4 4.6 5 4.4 4.9 ...
##  $ Sepal.Width : num  3.5 3 3.2 3.1 3.6 3.9 3.4 3.4 2.9 3.1 ...
##  $ Petal.Length: num  1.4 1.4 1.3 1.5 1.4 1.7 1.4 1.5 1.4 1.5 ...
##  $ Petal.Width : num  0.2 0.2 0.2 0.2 0.2 0.4 0.3 0.2 0.2 0.1 ...
##  $ Species     : Factor w/ 3 levels "setosa","versicolor",..: 1 1 1 1 1 1 1 1 1 1 ...

Based on user’s inputs, the application trains the model, applies it to test set and displays the model accuracy along with variable importance plot. Here is a sample run:

library(caret)
data(iris)

inTrain <- createDataPartition(iris$Species, p = 0.7, list = FALSE)
training <- iris[inTrain,]
testing <- iris [-inTrain,]

set.seed(1432)

trCtrl <- trainControl(method = "cv", number = 5)
mf <- train(Species ~ ., method = "rf", trControl = trCtrl, data = training)

prd <- predict(mf, testing)
cm <- confusionMatrix(prd, testing$Species)

print(paste("Model Accuracy:", round(cm$overall['Accuracy'] * 100, 2), '%'))

## [1] "Model Accuracy: 95.56 %"

User Selection Option

The user can select from the following options:

• Selecting Partition value for training and test datasets

Default: 0.7 (70%: 70% Training, 30% Testing)

• Predictor Variables Selection

Default: All 4. Please note that at least one predictor should be selected otherwise an error will be thrown.

• Resampling Method

Default: Not Selected. If selected, there are 4 options includes:

Bootstrapping, Cross Validation, Leave one out Cross Validation, None

• Number of Folds for cross validation/bootstrap

Default: 5

Output (Model Tab)

Based on user selection, the model gets trained and gets applied to testing Dataset. Following information is included:

• The Model

Please note the different parameters as they change based on user inputs.

• In-Sample Accuracy

When the model gets trained. Note that for method ‘none’, the accuracy is not displayed. Can you think of why?

• Confusion Matrix

When the model is applied to testing set

• Out-of-Sample Accuracy

When the model is applied to testing set

• Variable Importance Plot

It shows the relative importance of different predictors in the model.

Useful Links

• Project

– Shiny: https://promisinganuj.shinyapps.io/myNewApp/

– Github Repository: https://github.com/promisinganuj/ddp_assigment

It contains the “ui.R” and “shiny.R” files of the application.