Programming ML.NET

Chapter 6

Clustering Tasks

“It is only prudent never to place complete confidence in that by which we have even once been deceived.”

—René Descartes, Meditations on First Philosophy, 1641

There are many different approaches to classify data, each with its own pros and cons. ML.NET supports quite a few of these approaches and implements them in native trainers. In Chapter 5, we use logistic regression and the Stochastic Dual Coordinate Ascent (SDCA) algorithm to classify items, but other options, which sometimes are more effective, exist as well—specifically, the Support Vector Machine (SVM) algorithm. SVM and other aforementioned algorithms have one trait in common: they need to know the expected value to predict for each row being processed ...