Detecting Implicit Expressions of Affect from Text using Semantic Knowledge on Common Concept Properties

Detecting Implicit Expressions of Affect from Text using Semantic Knowledge on Common Concept Properties Alexandra Balahur, Hristo Tanev European Commission Joint Research Centre Institute for the Protection and Security of the Citizen Via E. Fermi 2749, 21027 Ispra (VA), Italy E-mail: {alexandra.balahur, hristo.tanev}@jrc.ec.europa.eu Abstract Emotions are an important part of the human experience. They are responsible for the adaptation and integration in the environment, offering, most of the time together with the cognitive system, the appropriate responses to stimuli in the environment. As such, they are an important component in decision-making processes. In today’s society, the avalanche of stimuli present in the environment (physical or virtual) makes people more prone to respond to stronger affective stimuli (i.e., those that are related to their basic needs and motivations – survival, food, shelter, etc.). In media reporting, this is translated in the use of arguments (factual data) that are known to trigger specific (strong, affective) behavioural reactions from the readers. This paper describes initial efforts to detect such arguments from text, based on the properties of concepts. The final system able to retrieve and label this type of data from the news in traditional and social platforms is intended to be integrated Europe Media Monitor family of applications to detect texts that trigger certain (especially negative) reactions from the public, with consequences on citizen safety and security. Keywords: emotion detection, media monitoring, argument mining, emotion triggers, common-sense knowledge, decision support 1. Introduction Paper Anger, fear, sadness, disgust, happiness or surprise. Any human being is able to relate to these emotions and give examples of situations when they can be felt and the possible manners in which they can be expressed. However, the mechanisms of emotion are in no way simple or straightforward to understand, explain or mimic. Until recently, emotions were considered separate from cognition (Forgas, 2008). The latter was considered in relation to rational thinking, reasoning, rigor, while affective phenomena were thought to “overwhelm or subvert rational mental processes” (Elster, 1985). Recent studies have shown that emotion and cognition actually go hand in hand, regulating our reactions and ensuring our well-being and balance, being an important part of the decision-making process (Forgas, 2008). Given the importance of these aspects, emotions have been studied intensely in the past years. In Natural Language Processing (NLP), the task that deals with the detection and classification of texts according to the polarity of the opinions they express is called Sentiment Analysis (SA). In spite of the dynamics of the field and the high amount of research that has been done under its umbrella in the past decade, systems dealing with sentiment analysis have only reached performance levels that are around 80% accuracy at most (reaching almost 90% in the systems that are employing deep learning and that are applied to product review classification, a more straightforward task). Of course, these levels are influenced by the types of text, the language and the language style. Apart from the inherent difficulties posed by the need for linguistic …

Content truncated to 500 words