The vast and ever-increasing text posting in "social networks" such as Facebook and Twitter, during the last 15 years, has produced an immense and rich text repository for several areas of knowledge. Therefore, text mining has recently become a very active and attractive area of research in computer science. The limited current understanding of the knowledge represented in these repositories has already influenced decision making for several companies, public figures and government agencies. However, the demand for full understanding and manipulation of these knowledge repositories has placed additional expectations on text mining. Hence, some algorithms have being developed for retrieving information, and most of them based on word recognition. This article discusses the use of phonetic algorithms to improve word recognition and proposes a precision measurement method to evaluate the results. It also presents some examples of applying some well-known phonetic algorithms on Spanish written texts using natural language processing techniques.

Keywords: Natural Language Processing; text processing; phonetic algorithms; Spanish text processing.

Article to be sent to Computational Intelligence (CSCI-ISCI)