A comprehensive review and analysis of supervised-learning and soft computing techniques for stress diagnosis in humans

Highlights

• A systematic review approach has been employed to study the prevalence and diagnosis of stress using SL and SC techniques.

• Different categories of stress, as well as its sources, symptoms, and consequences on human life are identified.

• The use of supervised learning techniques in diagnosis of human stress has been explained.

• The role of soft computing techniques in diagnosis of human stress is examined.

• The publishing trend for the related studies has been extensively examined and is presented.

Abstract

Stress is the most prevailing and global psychological condition that inevitably disrupts the mood and behavior of individuals. Chronic stress may gravely affect the physical, mental, and social behavior of victims and consequently induce myriad critical human disorders. Herein, a review has been presented where supervised learning (SL) and soft computing (SC) techniques used in stress diagnosis have been meticulously investigated to highlight the contributions, strengths, and challenges faced in the implementation of these methods in stress diagnostic models. A three-tier review strategy comprising of manuscript selection, data synthesis, and data analysis was adopted. The issues in SL strategies and the potential possibility of using hybrid techniques in stress diagnosis have been intensively investigated. The strengths and weaknesses of different SL (Bayesian classifier, random forest, support vector machine, and nearest neighbours) and SC (fuzzy logic, nature-inspired, and deep learning) techniques have been presented to obtain clear insights into these optimization strategies. The effects of social, behavioral, and biological stresses have been highlighted. The psychological, biological, and behavioral responses to stress have also been briefly elucidated. The findings of the study confirmed that different types of data/signals (related to skin temperature, electro-dermal activity, blood circulation, heart rate, facial expressions, etc.) have been used in stress diagnosis. Moreover, there is a potential scope for using distinct nature-inspired computing techniques (Genetic Algorithm, Particle Swarm Optimization, Ant Colony Optimization, Whale Optimization Algorithm, Butterfly Optimization, Harris Hawks Optimizer, and Crow Search Algorithm) and deep learning techniques (Deep-Belief Network, Convolutional-Neural Network, and Recurrent-Neural Network) on multimodal data compiled using behavioral testing, electroencephalogram signals, finger temperature, respiration rate, pupil diameter, galvanic-skin-response, and blood pressure. Likewise, there is a wider scope to investigate the use of SL and SC techniques in stress diagnosis using distinct dimensions such as sentiment analysis, speech recognition, handwriting recognition, and facial expressions. Finally, a hybrid model based on distinct computational methods influenced by both SL and SC techniques, adaption, parameter tuning, and the use of chaos, levy, and Gaussian distribution may address exploration and exploitation issues. However, factors such as real-time data collection, bias, integrity, multi-dimensional data, and data privacy make it challenging to design precise and innovative stress diagnostic systems based on artificial intelligence.

Introduction

Disease diagnosis is a pre-eminent multidisciplinary research area. Physiological diseases such as cancer and diabetes have often been considered as the most critical and prevalent diseases in humans worldwide [1]. However, psychiatric and neuropsychiatric disorders are the most prevalent human diseases. Stress appears to be a predominant, persistent, and global human psychiatric disorder, and it inevitably transforms the state of an individual [2,3]. The word “stress” is etymologically rooted in the Latin word “stringere,” which refers to an intensified rate of tension or adversity [4,5]. Stress may even be described as a panoramic rejoinder of the human body to any command. It is considered as the perception of threats with consequent anxiety, malaise, emotional trauma, and difficulty in adjustment [6]. Stress diagnosis is an important task. Generally, the process of stress diagnosis is handled by psychiatrists, counselors, or persons who are experts in human psychology. Measuring the blood pressure, heart rate, electroencephalogram (EEG) signals, finger temperature, and observing behavioral data are some of the major measures that are extensively employed in stress diagnosis. The consequences of stress can be observed in performance degradation, family conflicts, mood disorders, anxiety, and depression.

The research until now perceived that diverse supervised learning (SL) techniques such as Bayesian networks, Support Vector Machine (SVM), Random Forest (RF), Decision Tree (DT), Neural Network (NN), k-nearest neighbours (k-NN), ZeroR (ZR), naïve Bayes (NB), AdaBoost (AB), and J48 have been engaged in determining the stress intensity among different individuals [[7], [8], [9]]. Different researchers have achieved different rates of stress classification based on SL methods and data. Different performance measures such as accuracy, precision, sensitivity, specificity, F1-score, and recall were enumerated and analyzed for different computing techniques. Furthermore, to overcome the time and cost issues of conventional computing techniques, Zadeh proposed the concept of soft computing (SC) techniques to solve real-world problems in a comparatively less time and cost than those of SL or hard computing approaches [10]. SC methods deal with imprecision and uncertainty, whereas the conventional computing techniques are based on analytical models [11]. In general, SC is an amalgamation of different techniques, such as Fuzzy Logic (FL), Deep Learning (DL), and Nature-inspired Computing (NIC). Similar to SL techniques, SC techniques have also been effectively used in the early prognosis of different human maladies.

The existing research witnessed that lots of review articles on disease diagnosis have been published [[12], [13], [14], [15], [16], [17], [18], [19], [20], [21]]. Some of the core works associated with stress diagnosis are presented in Table 1.

Table 1 presents the research conducted by different researchers in stress diagnosis. The existing literature has revealed that several researchers have published manuscripts related to stress diagnosis. However, no systematic study pertaining to the use of SL and SC techniques for stress diagnosis has been conducted.

Herein, a meta-analysis of SL and SC techniques used in the diagnosis of stress, one of the paramount human psychiatric disorders, is presented. Efforts have been made to become acquainted with stress and its ramifications on human beings. The ramifications of psychological, biological, and behavioral responses to stress in the human body have also been asserted. The attributes corresponding to parenting stress and its consequences are also briefly elucidated. The detailed stress-related publication trend and its diagnosis using the SL and SC techniques were examined. This work is valuable for investigators who want to discern the effects of SL and SC techniques in psychiatric human disorders.

The review methodology used for this meta-analysis is described in Section 2. The synthesis and analysis of the data related with this research work are presented in Section 3. The results and discussion are presented in Section 4. Finally, the conclusions of this review are discussed in Section 5.