Text analysis

Paola Cerchiello

INTRODUCTION

In this chapter, we address one of the most powerful and relatively recent trends in artificial intelligence. Text mining dates back to the 1990s as a research area offering a wide range of tools and methods to complement standard quantitative analysis in every branch of research: in particular finance, economics, psychology and criminology, as well as the social sciences in general.

We can summarise the several definitions of text mining offered by researchers and practitioners as follows: text mining entails all activities needed to transform a set of, potentially extremely large, textual documents into analysable numeric information in order to extract content, topics, words, facts, opinions and sentiment. Thus, the key message from such a definition is that through text mining we are able to transform words into numbers and thence into knowledge.

The rapid and pervasive growth of the Internet, putting it at the disposal of a large part of the world’s population, has produced an incredibly huge amount of unstructured information, most of it textual in origin. This has allowed the parallel development of text mining in response to the clear need to manage or