Machine Learning

This chapter explores machine learning. This topic is closely related to statistical modeling, which we considered in Chapter 14 because both use data to describe and predict outcomes of uncertain or unknown processes. The approach taken in statistical modeling emphasizes understanding how the data is generated by devising models that describe the underlying process behavior and fitting the model’s parameters to the observed data. If the model fits the data well and satisfies the relevant model assumptions, then the model can be assumed to give a holistic description of the process. Such a model can, for example, be used to compute statistics with known distributions and evaluate statistical tests. However, if the data is too complex to be explained using available statistical models, this approach has reached its limits. In machine learning, on the other hand, the process that generates the data and potential models thereof is not central. Instead, the observed data and the explanatory variables are the fundamental starting point of a machine-learning application. Given data, machine-learning methods can be used to find patterns and structures in the data, which can be used to predict the outcome of new observations. Machine learning, therefore, does not provide an understanding of how data is generated. Because fewer assumptions are made regarding the distribution and statistical properties of the data, we typically cannot compute statistics and perform statistical tests regarding the significance of specific observations. Instead, machine learning strongly emphasizes the accuracy with which new observations are predicted.