Negation Processing in Spanish and its Application to Sentiment Analysis

Natural Language Processing is the area of Artificial Intelligence that aims to develop computationally efficient mechanisms to facilitate communication between people and machines through natural language. To ensure that machines are capable of processing, understanding and generating human language, a wide range of linguistic phenomena must be taken into account, such as negation, irony or sarcasm, which are used to give words a different meaning. This doctoral thesis focuses on the study of negation, a complex linguistic phenomenon that we use in our daily communication. In contrast to most of the existing studies to date, it is carried out on Spanish texts, because i) it is the second language with most native speakers, ii) it is the third language most used on the Internet, and iii) there are no negation processing systems available on this language. Not even Google has been able to properly process Spanish negation. For example, if we perform the search <<películas que no sean de aventuras>>, we can see that it returns adventure films, when it should return films of other themes.

This thesis has been awarded as the best doctoral thesis in Natural Language Processing at the 36th International Conference of the Spanish Society for Natural Language Processing (SEPLN 2020).

Salud María Jiménez Zafra