A Survey of Machine Learning Methods for Network Planning

Abstract

In this chapter, machine learning, artificial intelligence, network planning, and the connection between these concepts are discussed in general terms. In order to generate new predictions, machine learning models attempt to exploit the fundamental relationships and patterns in your data. Image recognition serves as a common illustration. We are aware that the connections between image pixels and labels (such "dogs" or "cats") are intricate. It is highly challenging to find this intricate relationship using an equation that takes the image of a cat or a dog as input, and it constructs a new equation for each label, making it a scalable solution. Instead, machine learning techniques look for situations where the pixels and labels don't match. The goal of machine learning is to identify if an invisible image depicts a dog or a cat by removing probable patterns using a training set that includes several instances of any type of cat or dog. Similar to this, machine learning techniques may eliminate the relationship between chemical systems and properties when given sufficient examples, as well as the ability to otherwise dissolve extremely complex equations.