A hybrid approach for word emotion recognition

Abstract

In recent years, with plenty of online resources constantly emerging, emotion recognition in text has become increasingly important in human-computer interaction. Word emotion plays a very important role in emotion analysis of sentences or documents. This paper proposes a hybrid approach to recognition of word emotion in the dimension of eight emotion categories with corresponding intensities based on the Chinese emotion corpus. First, we present a new algorithm of semantic similarity computation for aiding emotion intensity computation and design a new algorithm of emotion vector computation by making use of both morpheme characteristics and semantic relations. And then, we adopt support vector machine model for the secondary classification to the words whose emotions cannot be calculated by the semantic analysis algorithm. Our approach achieves the accuracy of 54.00% and 78.75% for exact match and all five types of hit, respectively, on the basis of the core emotion lexicon CL4. Experimental results show that the integration of morpheme characteristics and semantic relations can improve the classification accuracy efficiently.

Trends detection of flu based on ensemble models with emotional factors from social networks

“…In order to make comparisons, we also revised the CRFs to detect the flu state of a single person and construct a supervised flu state detection model . The posts of a flu infected person can be treated as a sequence, and the flu state detection of the person can be treated as a sequence labeling problem, which can be solved by CRFs .…”

TFSM‐based dialogue management model framework for affective dialogue systems

“…So how to endow the machine with the ability of emotional interaction with humans has become an important and challenging subject in intelligent human-computer interaction [17]. To date, most studies in affective computing are focused on detecting and recognizing emotional information with different modalities [18,19], such as speech, facial expression, posture, text, physiological information [20][21][22], etc. Affective dialogue system can be viewed as an enhanced version of SDS that incorporates emotion recognition, emotional interaction, as well as emotion generation and expression ( Fig.…”

PI1710-3-EmoAnalysi : desarrollo de una herramienta software para el análisis de emociones presentes en la transcripción de llamadas telefónicas para apoyar procesos de inteligencia de negocios de un centro de atención a clientes

“…Las emociones definidas para este modelo, fueron estudiadas y analizadas dentro de un grupo de individuos interculturales, donde los individuos identificaron con un alto grado de fiabilidad las expresiones emocionales al observarlas en fotografías. Las emociones que se definieron en el año 1972 son: tristeza, alegría, ira, miedo, disgusto y sorpresa [13].…”

“…Posteriormente se adopta el modelo Support Vector Machine (SVM) para la clasificación secundaria de las palabras cuyas emociones no fueron calculadas por el primer algoritmo. Este enfoque tiene una exactitud de aproximadamente 54% [13].…”

“…SentiWordNet es un recurso léxico tos, donde a cada una de las vo, negativo y neutro [13] La versión del SentiWordNet descargada corresponde a 3.0, que valida la versión del WordNet 3.0, razón por la cual p Affect. Adicionalmente, debido a que se encuentra en inglés, se requirió de la traducción y validación de esta información.…”