February 1, 2018 (Vol. 38, No. 3)
Strong Points: Clear visualization of different algorithms, easy to use
Weak Points: Little background provided, few algorithms
Machine learning is becoming more prevalent in biology. As such, it will be increasingly useful for biologists to familiarize themselves with some of the basic methods and approaches of the machine-learning field, such as K-means clustering and principle component analysis (PCA). The Deepvis app provides a very promising (albeit somewhat inchoate) resource to gain an intuitive understanding of analyses like these through visual depictions. In addition to K-means clustering and PCA, the app currently includes two other machine-learning algorithms: hierarchical clustering analysis (HCA) and linear discriminant analysis (LDA). App users select one of these from the home screen and are then taken to a blank screen onto which they can add their own data nodes in whatever pattern they choose. The app will then highlight how those points would be clustered (into either two or three clusters) based on the corresponding algorithm. The app links to the corresponding Wikipedia pages as the means to provide background information.