سیستم ارزیابی احساسات مبتنی بر الکتروانسفالوگرام با استفاده از تکنیک های داده کاوی و هستی شناسی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|46674||2015||12 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Applied Soft Computing, Volume 30, May 2015, Pages 663–674
Currently, emotion is considered as a critical aspect of human behavior; thus it should be embedded within the reasoning module in an intelligent system where the aim is to anticipate or respond to human reactions. Therefore, current research in data mining shows an increasing interest in emotion assessment for improving human–machine interaction. Based on the analysis of electroencephalogram (EEG) which derives from automatic nervous system responses, computers can assess user emotions and find correlations between significant EEG features extracted from the raw data and the human emotional states. With the advent of modern signal processing techniques, the evaluative power of human emotion derived from EEG is increased exponentially due to the huge number of features that are typically extracted from the EEG signals. Notwithstanding that the expanded set of features could allow computers to evaluate emotions in an accurate way, it is too complex a task to manage in a structured way and, for the reasons stated, methods and approaches to enable both EEG information management and evaluation are necessary to support emotion assessment. Starting from this consideration, this paper proposes an enhanced EEG-based emotion assessment system exploiting a collection of ontological models representing EEG feature sets and arousal–valence space (two-dimensional emotion scale), statistical tests capable of evaluating the gender-specific correlations between EEG features and emotional states, and a classification methodology inferring arousal and valence levels. As will be shown in the experimental section where the proposed approach has been tested on a public dataset, the experimental results demonstrate that better performance in emotion assessment can be achieved using our framework as compared with other studies using the same dataset and with three other classification techniques.