Estimating the pattern frequency spectrum inside the browser

We present a browser application for estimating the number of frequent patterns, in particular itemsets, as well as the pattern frequency spectrum. The pattern frequency spectrum is defined as the function that shows for every value of the frequency threshold σ the number of patterns that are frequent in a given dataset. Our demo implements a recent algorithm proposed by the authors for finding the spectrum. The application is 100% JavaScript, and runs in all modern browsers. We observe that modern JavaScript engines can deliver performance that makes it viable to run non-trivial data analysis algorithms in browser applications. Keywords—pattern mining, JavaScript, web browser