Introduction

1.1 Introduction

Following a drizzling, we take a walk on the wet street. Feeling the gentle breeze and seeing the sunset glow, we bet the weather must be nice tomorrow. Walking to a fruit stand, we pick up a green watermelon with curly root and muffled sound; while hoping the watermelon is ripe, we also expect some good aca- demic marks this semester after all the hard work on studies. We wish readers to share the same confidence in their studies, but to begin with, let us take an informal discussion on what is machine learning .

Taking a closer look at the scenario described above, we notice that it involves many experience-based predictions. For example, why would we expect beautiful weather tomorrow after observing the gentle breeze and sunset glow? We expect this beautiful weather because,from our experience,theweather on the following day is often beautiful when we experience such a scene in the present day. Also, why do we pick the watermelon with green color, curly root, and muffled sound? It is because we have eaten and enjoyed many watermelons, and those sat- isfying the above criteria are usually ripe. Similarly, our learn- ing experience tells us that hard work leads to good academic marks. We are confident in our predictions because we learned from experience and made experience-based decisions.

Mitchell ( 1997 ) provides a more formal definition: ‘‘A computer program is said to learn from experience $E$ for some class of tasks $T$ and performance measure $P$ , if its performance at tasks in $T$ , as measured by $P$ , improves with experience $E$ .’’

E.g., Hand et al. ( 2001 ).

While humans learn from experience, can computers do the same? The answer is ‘‘yes’’, and machine learning is what we need. Machine learning is the technique that improves system performance by learning from experience via computational methods. In computer systems, experience exists in the form of data, and the main task of machine learning is to develop learning algorithms that build models from data. By feeding the learning algorithm with experience data, we obtain a model that can make predictions (e.g., the watermelon is ripe) on new observations (e.g., an uncut watermelon). If we consider com- puter science as the subject of algorithms, then machine learn- ing is the subject of learning algorithms .

In this book, we use ‘‘model’’ as a general term for the out- come learned from data. In some other literature, the term ‘‘model’’may refer to the global outcome (e.g., a decision tree), while the term ‘‘pattern’’ refers to the local outcome (e.g., a single rule).