Mining Big Data With Computational Methods
Annie Waldherr, Gerhard Heyer, Patrick Jähnichen, Andreas Niekler, Gregor Wiedemann · 2015
This chapter discusses on the state of the art in text mining methods, discuss potential applications and limitations for political communication research. The age of big data poses enormous challenges to traditional methods of empirical research who deals with a seemingly endless amount of web sources. The broad set of methods to semantically structure, large amounts of unstructured text data is referred to as text mining. Natural language processing (NLP) text mining methods allow the automatic capture of the semantics of texts in unstructured corpora. In NLP, three types of semantic processing models: patterns of character strings, logical representations of entity relations, and distributional semantics. The big advantage of text mining is that we gain an overview on the content of vast text corpora with limited efforts. Text mining proves to be helpful for identifying relevant text documents from large data bases to prepare an in-depth manual content analysis.