Article Preview
TopIntroduction
Human emotions get influenced by a variety of circumstantial and psychological factors including personality. Identifying, modeling, and synthesizing these emotions are the key contributions of Affective computing especially in case of psychological ill patients. Affect (emotion) has a good relation with the personality of a human being (Subramanian et al., 2018). Personality refers to the features of an individual that derive their behavior pattern which assists in judging them in the various life outcomes like family relationships, career success, mental health problems, etc (Miranda-Correa et al., 2018). Any particular emotion, personal memories, daily routine lifestyle, and social relationships play an important role in affecting the personality of an individual. Researchers have proposed many theories and models of personality, however, the most recognition was gained by the Big Five-factor model that plays a unique role in the scientific community (Zhao et al., 2017). To identify the personality of healthy subjects, multiple tools and techniques including sensors, physiological signals acquisition, processing, etc. have been explored (Miranda-Correa et al., 2018; Subramanian et al., 2018). Other than this, for mentally unhealthy patients, sensing systems and questionnaire-based analysis were done to judge their personality. Sensing technologies include fetching signals by placing various electrodes/sensors on the human body and they are the latest and accurate methods as they directly measure the activities of the internal nervous system that is uncontrollable by the person and give spontaneous results (Garcia-Ceja et al., 2018).
In this paper, the literature on personality prediction for mentally healthy and unhealthy subjects is discussed in brief. As sensing-based systems play a crucial role in personality identification, thus to give readers a more understanding of the whole process, EEG signals of healthy subjects are chosen for the analysis as they are directly connected to the limbic system of the central nervous system. EEG signals are the best measure to analyze the brain signals and ensure the identification of personality traits as well (Kim & Jo, 2018). The basic processing flow of the physiological signal-based personality identification is shown in Figure 1. It helps future researchers to get an idea and apply the same for the mentally unhealthy patients in their respective domains. This process includes signal acquisition and its processing to remove noise and various artifacts. Following this feature extraction and at last, classification took place to achieve the results.
Figure 1. The processing flow of personality prediction using EEG signals
In brief, the aim of the current work is:
- •
To introduce physiological signals, machine learning algorithms as well as personality dimensions.
- •
To give knowledge about the work done by various researchers related to personality prediction.
- •
To provide information about the methodology used for personality prediction processes mainly for healthy subjects.
To accomplish the goal of the current work, this paper is structured in multiple sections. Section 2 explains the work related to different measures of personality prediction for healthy and mentally unhealthy subjects. After that, the whole processing flow for personality prediction using brain signals along with EEG and used dataset description is explained in Section 3. Following this, experimental results and the interrelation of personality dimensions are shown in Section 4. At last, a conclusion is drawn with some of the future challenges in Section 5.
TopBackground
This section presented background work done on (a) Personality prediction methods (b) Personality prediction of mentally unhealthy subjects, and (c) Personality prediction of healthy subjects. Further it also explores the methodology that defines the personality prediction process of healthy subjects which may later on can be used for unhealthy subjects treatment as well.