K-means versus Kernel Kmeans: Shiny Application

This shiny application is designed to show the difference between the K-means and Kernel K-means clustering methods.

Using a small non-linear data set, the user can select K-means or Kernel K-means algorithms and define the number of clusters and $$\sigma$$ (when using Kernel K-means).

The resulting output with be a plot above the user interface showing the clusters and their centers. It will also run a cross-validation check on a truth file when the cluster number is set to 2. If it is above two it will generate a random vector using sample(input$k,nrow(data),rep=TRUE) as the generator in the user file.

In the future, I aim to add more features to this app, including:

• Selectable datasets with truth files
• Different algorithms to choose from
• Plot output options
• More...

There is a slider presentation available at this Github page and the Shiny Application is available at this page.