ABSTRACT
In this chapter, we discuss the well-known data science tool called “neural network regression,” from the point of view of means of conditional distributions. The chapter begins defining “universal approximators,” and notes that, like polynomial and Fourier series functions, the neural network function is another universal approximator. Examples are given to understand the use and meaning of the neural network function, using both noiseless and noisy data. Examples comparing neural network functions with polynomial functions are also given.
