The results and findings section presents the outcomes of the systematic mapping study. It provides a comprehensive overview of the identified computer-based tools for emotion and personality identification, highlighting their methodologies, effectiveness, and application domains.
Study Participants
Understanding the composition and characteristics of study participants is essential for evaluating the relevance and robustness of research findings in the domain of tools designed to identify emotions and personality traits within computer science. In this section, we provide insights into the diverse range of participants involved in the studies we selected during our systematic mapping study, along with the corresponding participant numbers. The demographic information of the study participants, such as age, gender, education level, and cultural background, plays a crucial role in understanding the nuances of emotional and personality traits, as shown in Table 6. These factors can significantly influence how individuals express emotions and manifest their personality traits in digital interactions. Our selected studies encompassed a diverse range of participant demographics, with participant numbers ranging from small cohorts of 20 to larger, more extensive samples exceeding 200[56] [57].
The size and diversity of the study sample are key considerations for the generalizability of the research findings. A larger and more diverse sample size enhances the external validity of the study results. In our selected studies, we observed variations in sample sizes. Some studies featured relatively small cohorts with approximately 30 participants, while others boasted larger, more representative samples, with participant numbers exceeding 200. Additionally, researchers often aim to include participants from various cultural backgrounds, ethnicities, and socioeconomic statuses to capture a broader range of emotional and personality variations [58]. User characteristics encompass a variety of factors, including prior experience with technology, familiarity with the specific application or tool being studied, and individual differences in cognitive and emotional traits. These characteristics can significantly influence how participants engage with technology and how their emotional responses and personality traits manifest during digital interactions. Our selected studies considered a range of user characteristics, with participant numbers correlating to the specific characteristics under investigation [59] [60].
Ethical considerations are of paramount importance in studies involving human participants. Ensuring ethical practices, including obtaining informed consent from participants, protecting their privacy and data, and minimizing potential harm, is crucial. The studies we selected, with their respective participant numbers, adhered to ethical guidelines and standards, reflecting a commitment to upholding the rights and well-being of the participants involved in the research [61]. Understanding the characteristics, demographics, and participant numbers of study participants is integral to interpreting and contextualizing research findings. It enables researchers and practitioners to assess the applicability of results to diverse user populations, ensuring that insights drawn from these studies are both meaningful and broadly relevant to the field of computer science.
Understanding the characteristics, demographics, and participant numbers of study participants is integral to interpreting and contextualizing research findings. It enables researchers and practitioners to assess the applicability of results to diverse user populations, ensuring that insights drawn from these studies are both meaningful and broadly relevant to the field of computer science.
Table 6
Summary of Study Participants and Study Characteristics
Aspect
|
Description
|
Study Participants
|
-Range of participants with numbers from 20 to over 200 [56] [57].
|
Sample Size
|
-Varied, from approximately 30 to over 200, aiming for diverse demographics [58].
|
User Characteristics
|
- Factors include prior tech experience and individual traits [59] [60].
|
Ethical Considerations
|
- Crucial adherence to guidelines, including consent and privacy protection [61].
|
Table 7 presents a summary of various computerized tools developed for the assessment of emotions and personality traits. In the rapidly evolving landscape of mental health and psychological research, these innovative tools utilize a range of technologies, from natural language processing to facial recognition and physiological monitoring. Each tool serves a distinct purpose, targeting specific populations and age groups while addressing various mental health concerns. The table offers a snapshot of these tools, their applications, and the key psychometric properties assessed to evaluate their effectiveness in enhancing our understanding of emotional and personality characteristics.
Table 7
Summary of computerized tools for emotion and personality assessment
Author/Year
|
Computerized tools proposed/used
|
Description of tool
|
The population of study e.g. students, university students, women, etc.
|
Age of study participants
|
Study sample size
|
Methodology
|
Type of mental health concern
|
Psychometric properties of tool assessed
|
Smirnov et al. (2021)[62]
|
Titanis
|
A tool for intelligent text analysis in social media
|
English-speaking social media users
|
18–65 years
|
1000
|
Cross-sectional
|
Depression, anxiety, and stress
|
Validity and Reliability
|
De et al. (2015)[63]
|
Eigenface approach
|
A facial expression recognition algorithm based on the eigenface approach
|
Indian university students
|
18–25 years
|
100
|
Machine Learning
|
Anxiety
|
Accuracy and Sensitivity
|
Exler et al. (2016)[64]
|
Wearable system for mood assessment
|
A system that uses smartphone features and data from mobile ECGs to assess mood
|
German adults
|
22–46 years
|
6
|
Mobile App Data Analysis
|
Stress
|
Validity and User Experience
|
Ovur et al. (2021)[65]
|
Autonomous learning framework for sEMG-based hand gesture recognition
|
A machine learning framework that uses surface electromyography (sEMG) and depth data to recognize hand gestures
|
Healthy adults
|
22–30 years
|
10
|
Machine Learning and Gesture Analysis
|
None
|
Precision and Recall
|
Azam et al. (2019)[66]
|
Smartphone-based mindful breathing app
|
A smartphone-based app that guides users through mindful breathing exercises
|
University students
|
18–25 years
|
100
|
Randomized controlled trial
|
Depression and anxiety
|
Reliability and Sensitivity
|
Nave et al. (2018)[67]
|
Social media mobile photography
|
Using social media mobile photography to self-track emotional states
|
General Smartphone Users
|
18–25 years
|
30
|
Mobile App Data Collection
|
Emotional Well-being
|
Test-retest Reliability
|
Villatoro-Tello et al. (2021)[68]
|
Clinical interviews as a support tool for depression detection
|
Speech analysis tool for emotion detection in clinical interviews
|
Adults
|
18–65 years
|
100
|
Machine learning
|
Mood Disorders
|
Sensitivity and Specificity
|
Cheong et al. (2020)[69]
|
Wearable sensor-based system
|
Physiological sensors for monitoring mood changes in elderly patients
|
Elderly individuals
|
65 + years
|
20
|
Physiological Monitoring
|
Mood Changes in the Elderly
|
Sensitivity and Specificity
|
Fadhil et al. (2019)[70]
|
CoachAI
|
A conversational agent assisted health coaching platform
|
General Smartphone Users
|
18–25 years
|
20
|
Natural Language Processing (NLP)
|
Emotional Well-being
|
Validity and User Experience
|
Metin et al. (2022)[71]
|
Deep learning method
|
A deep learning method that uses EEG data to differentiate patients with bipolar disorder from controls
|
Patients with Bipolar Disorder
|
18–65 years
|
100
|
Electroencephalography (EEG)
|
Bipolar Disorder
|
Test-retest Reliability
|
Cuijpers et al. (2016)[72]
|
Internet-based cognitive behavioral therapy (iCBT)
|
A self-guided, computer-based program that teaches CBT skills for depression
|
Adults with depression
|
18–65 years old
|
242
|
Randomized controlled trial
|
Depression
|
Efficacy, acceptability, and usability
|
Andrews et al. (2018)[73]
|
Virtual reality exposure therapy (VRET)
|
A computer-generated simulation that exposes people to feared situations in a safe and controlled environment
|
Adults with anxiety disorders
|
18–65 years old
|
100
|
Randomized controlled trial
|
Anxiety disorders
|
Efficacy, acceptability, and usability
|
Wolf et al. (2016)[74]
|
Mobile phone intervention
|
A mobile phone intervention that provides guided mindfulness exercises to people with depressive symptoms
|
Adults with depression
|
18–65 years old
|
50
|
Cross-sectional
|
Depression
|
Efficacy, acceptability, and usability
|
Overall, the results and findings provide a comprehensive synthesis of the current state of research on computer-based tools employing machine learning techniques for emotion and personality identification. This section highlights the strengths and limitations of the identified tools, guiding researchers and practitioners in developing more accurate and reliable means of understanding human emotions and personality traits.
RQ1: What are the existing challenges of developing tools for identifying emotions and personality traits?
Identifying emotions and personality traits using computer-based tools is a burgeoning field with numerous challenges. These challenges span technological, ethical, and psychological domains, reflecting the complexity of capturing and interpreting human emotions and personality traits through digital means. Below, we delve into the existing challenges faced in the development of such tools, supported by relevant references.
Obtaining high-quality and diverse datasets for training machine learning models is a fundamental challenge. Biases in training data, limited sample sizes, and data privacy concerns can hinder the development of accurate emotion and personality detection tools. Emotions and personality are multidimensional constructs, and integrating insights from psychology, neuroscience, and computer science is challenging. Ensuring that tools accurately capture these complex aspects of human behavior is an ongoing struggle. Emotions and personality traits can vary across cultures, leading to the challenge of developing tools that are culturally sensitive and applicable to a global audience [75] [76].
Achieving real-time emotion detection and personality assessment remains a challenge, particularly in contexts where timely responses are critical, such as mental health applications [77] [78]. Gathering sensitive emotional and personality data raises significant ethical questions. Safeguarding user privacy, obtaining informed consent, and protecting against data breaches are essential but challenging aspects of tool development [79] [80]. Emotions are inherently subjective, making it difficult to create objective and universally applicable measurement tools. The challenge lies in developing methods that can reliably interpret subjective emotional experiences [81]. Integrating data from various sources, such as text, facial expressions, physiological signals, and voice, into a cohesive assessment of emotions and personality traits is a complex challenge requiring advanced multimodal fusion techniques [82]. Ensuring the validity and generalizability of tools across diverse populations and contexts is a continual challenge. Validation methodologies need to evolve to encompass the intricacies of emotions and personality [83]. The development of tools that users find acceptable and easy to use is vital for adoption. The challenge is in designing interfaces and interactions that are user-friendly and nonintrusive [84]. Machine learning models trained on biased data can perpetuate stereotypes and inequalities. Ensuring fairness in emotion and personality assessment tools is an ongoing challenge [85].
The development of tools for identifying emotions and personality traits is a multifaceted endeavor that involves technological, ethical, and psychological complexities. Addressing these challenges requires collaboration across disciplines, rigorous data collection and analysis, ethical considerations, and a commitment to ongoing improvement in tool design and validation. The following table summarizes the existing challenges of developing tools for identifying emotions and personality traits.
Table 8
Challenges in Developing Tools for Identifying Emotions and Personality Traits
Challenge
|
Description
|
References
|
Obtaining high-quality and diverse datasets for training machine learning models
|
Biases in training data, limited sample sizes, and data privacy concerns can hinder the development of accurate emotion and personality detection tools.
|
[75], [76]
|
Integrating insights from psychology, neuroscience, and computer science
|
Emotions and personality traits are multidimensional constructs, and integrating insights from various disciplines is challenging.
|
[75], [76]
|
Cultural sensitivity and global applicability
|
Emotions and personality traits can vary across cultures, leading to the challenge of developing tools that are culturally sensitive and applicable to a global audience.
|
[75], [76]
|
Real-time emotion detection and personality assessment
|
Achieving real-time detection and assessment, particularly in critical contexts like mental health applications, remains challenging.
|
[77], [78]
|
Ethical considerations and data privacy
|
Safeguarding user privacy, obtaining informed consent, and protecting against data breaches are essential but challenging aspects of tool development.
|
[79], [80]
|
Subjectivity of emotions
|
Emotions are inherently subjective, making it difficult to create objective and universally applicable measurement tools.
|
[81]
|
Multimodal fusion techniques
|
Integrating data from various sources into a cohesive assessment of emotions and personality traits requires advanced multimodal fusion techniques.
|
[82]
|
Validity and generalizability across diverse populations and contexts
|
Ensuring the validity and generalizability of tools across diverse populations and contexts is a continual challenge.
|
[83]
|
User acceptance and usability
|
Developing tools that users find acceptable and easy to use is vital for adoption. The challenge is in designing user-friendly and nonintrusive interfaces and interactions.
|
[84]
|
Fairness in machine learning models
|
Machine learning models trained on biased data can perpetuate stereotypes and inequalities. Ensuring fairness in emotion and personality assessment tools is an ongoing challenge.
|
85]
|
Table 8 provides a concise overview of the challenges encountered in the development of tools for identifying emotions and personality traits, along with relevant references supporting each challenge.
RQ1.1 What are the data collection methods used in identifying emotions and personality traits?
In studies aimed at identifying emotions and personality traits, various data collection methods are employed to gather the necessary information for analysis. These methods are chosen based on the research objectives, the nature of the traits being studied, and the available resources. There are several common data collection methods used in such studies.
Surveys and questionnaires are widely used to collect self-reported data on personality traits and emotional experiences. Participants answered a series of standardized questions designed to assess their personality characteristics or emotional states. Examples include the Big Five Inventory for Personality Traits and the Positive and Negative Affect Schedule (PANAS) for emotions [26] [27]. Interviews, including structured, semistructured, and open-ended formats, allow researchers to gather in-depth information about an individual's emotions and personality. Clinical interviews, for instance, are used in psychological assessments to diagnose personality disorders [62] [68] [71].
Behavioral observations involve systematically watching and recording an individual's actions, facial expressions, body language, and verbal cues to infer their emotional states and personality traits. This method is often used in clinical and research settings. Physiological data collection methods, such as electrocardiography (ECG), electroencephalography (EEG), and functional magnetic resonance imaging (fMRI), are used to monitor physiological responses associated with emotions [64] [65] [71] [72]. These measures can provide objective data on emotional reactions. Textual data sources, such as social media posts, emails, or written essays, can be analyzed using natural language processing (NLP) techniques to extract emotional content and personality traits. Sentiment analysis and linguistic analysis are common approaches. Facial expression analysis involves using computer vision techniques to recognize and analyze facial expressions in images or videos [71]. Facial expression analysis is a valuable method for assessing emotions noninvasively.
Experimental tasks and games are designed to elicit specific emotional responses or behaviors from participants. These tasks are often used in psychology and neuroscience studies to study emotions in controlled settings. Wearable sensors, such as heart rate monitors and skin conductance sensors, can capture physiological changes associated with emotions in real time [69]. These sensors are used in both research and clinical applications. Mobile applications and smart devices with built-in sensors, such as accelerometers and GPSs, can collect data on users' behaviors, movements, and locations, which can be indicative of emotions and personality traits. Biometric data, including fingerprints, iris scans, and voice recordings, can be used to identify unique characteristics related to personality traits and emotional states [66] [67] [74].
The choice of data collection method depends on the specific research goals, the target population, and the feasibility of the approach. Many studies employ a combination of these methods to obtain a comprehensive understanding of emotions and personality traits. In Table 7, we outline and discuss the diverse array of data collection methods commonly utilized in studies aimed at identifying emotions and personality traits. Each method offers unique insights and advantages, contributing to a comprehensive understanding of human behavior and psychological processes.
Table 9
Data collection methods for identifying emotions and personality traits
Data Collection Method
|
Description
|
Example Studies
|
Surveys and Questionnaires
|
Participants answer standardized questions to assess personality traits or emotional states.
|
[26], [27]
|
Interviews
|
Structured or semistructured interviews gather in-depth information about emotions and personality.
|
[62], [68], [71]
|
Behavioral Observations
|
Systematic observation of actions, facial expressions, and verbal cues to infer emotions and traits.
|
[71]
|
Physiological Data Collection
|
Monitoring physiological responses such as ECG, EEG, and fMRI to assess emotions.
|
[64], [65], [71], [72]
|
Textual Data Analysis
|
Analyzing social media posts or written text to extract emotional content and personality traits using NLP.
|
[71]
|
Facial Expression Analysis
|
Using computer vision techniques to recognize and analyze facial expressions in images or video.
|
[71]
|
Experimental Tasks and Games
|
Designed tasks to elicit specific emotional responses in controlled settings.
|
[69]
|
Wearable Sensors
|
Using sensors like heart rate monitors to capture physiological changes associated with emotions.
|
[69]
|
Mobile Applications and Smart Devices
|
Collecting behavioral data, movements, and locations from mobile apps and smart devices.
|
[66], [67], [74]
|
Biometric Data
|
Utilizing biometric data like fingerprints and voice recordings to identify unique characteristics related to emotions and traits.
|
[66], [67]
|
The methods shown in Table 9 offer various ways to gather data on emotions and personality traits, providing researchers with diverse approaches to studying these phenomena.
RQ1.2 What are the data analysis methods used in identifying emotions and personality traits studies?
Identifying emotions and personality traits in studies involves the use of various data analysis methods to process and make sense of the collected data. The choice of analysis methods depends on the nature of the data, the research objectives, and the complexity of the traits being studied. There are several common data analysis methods used in such studies.
Descriptive statistics, such as the mean, median, and standard deviation, are used to summarize and describe the central tendencies and distributions of personality trait scores or emotional responses obtained from surveys or questionnaires [26]. Factor analysis is employed to identify underlying factors or dimensions within a set of observed variables. In personality research, the Big Five personality traits (i.e., openness, conscientiousness, extraversion, agreeableness, neuroticism) are often extracted from a pool of related questionnaire items [27].
Cluster analysis groups individuals with similar personality profiles or emotional responses into clusters or categories. It can help individuals identify distinct personality types or emotional patterns within a population [62]. Regression analysis is used to explore relationships between personality traits or emotional states and other variables. For example, it can be used to examine how personality traits predict specific behaviors or outcomes [71].
Machine learning techniques, including classification and regression algorithms, are increasingly used to predict and classify emotions and personality traits based on various data sources. Support vector machines, decision trees, and neural networks are common choices [63] [65]. NLP methods are employed to analyze textual data, such as social media posts or written content, to extract emotional content, sentiment, or personality traits. Techniques include sentiment analysis and text classification [71].
Content analysis involves systematically coding and categorizing qualitative data, such as interview transcripts or open-ended survey responses, to identify themes and patterns related to emotions and personality [64] [65]. In studies that involve neuroimaging data (e.g., fMRI, EEG), sophisticated image analysis techniques are used to identify brain regions associated with specific emotions or personality traits [71] [72].
In facial expression analysis, machine vision algorithms are used to detect and classify facial expressions, helping to identify emotional states based on facial cues [63]. Time series analysis is applied when studying how emotions or personality traits change over time. It can reveal temporal patterns and trends in emotional responses [65]. SEM is used to test complex models that involve multiple variables and relationships. It can be used to examine how personality traits interact and influence outcomes [69].
Biometric data, such as heart rate variability or skin conductance, are analyzed to identify physiological patterns associated with specific emotional states. Data visualization techniques, including charts, graphs, and heatmaps, are used to present and interpret complex data related to emotions and personality traits [71].
The choice of data analysis methods depends on the research goals and the type of data collected. Many studies combine multiple methods to gain a comprehensive understanding of emotions and personality traits, especially when using diverse data sources such as surveys, physiological measures, and textual data.
The following table summarizes the data analysis methods used in identifying emotions and personality trait studies:
Table 10
Summary of Data Analysis Methods for Identifying Emotions and Personality Traits
Data Analysis Method
|
Description
|
Descriptive Statistics
|
Summarizes central tendencies and distributions of personality trait scores or emotional responses obtained from surveys or questionnaires.
|
Factor Analysis
|
Identifies underlying factors or dimensions within observed variables, often used to extract the Big Five personality traits from related questionnaire items.
|
Cluster Analysis
|
Groups individuals with similar personality profiles or emotional responses into clusters or categories, aiding in identifying distinct personality types.
|
Regression Analysis
|
Explores relationships between personality traits or emotional states and other variables, predicting specific behaviors or outcomes.
|
Machine Learning Techniques
|
Predicts and classifies emotions and personality traits based on various data sources, including support vector machines, decision trees, and neural networks.
|
Natural Language Processing
|
Analyzes textual data, such as social media posts or written content, extracting emotional content, sentiment, or personality traits using sentiment analysis.
|
Content Analysis
|
Systematically codes and categorizes qualitative data, such as interview transcripts or open-ended survey responses, identifying themes related to emotions.
|
Neuroimaging Analysis
|
Utilizes sophisticated image analysis techniques to identify brain regions associated with specific emotions or personality traits.
|
Facial Expression Analysis
|
Employs machine vision algorithms to detect and classify facial expressions, aiding in identifying emotional states based on facial cues.
|
Time Series Analysis
|
Examines how emotions or personality traits change over time, revealing temporal patterns and trends in emotional responses.
|
Structural Equation Modeling
|
Tests complex models involving multiple variables and relationships, examining how personality traits interact and influence outcomes.
|
Biometric Data Analysis
|
Analyzes physiological patterns associated with specific emotional states, including heart rate variability or skin conductance.
|
Data Visualization Techniques
|
Presents and interprets complex data related to emotions and personality traits using charts, graphs, and heatmaps.
|
Table 10 provides an overview of the various data analysis methods employed in studies focused on identifying emotions and personality traits, highlighting the diverse approaches used to gain insights into human behavior and psychological processes.
RQ2 - What are the key open issues in developing tools for identifying emotions and personality traits?
Identifying emotions and personality traits through tools and technologies is an evolving field, and several key open issues persist. These challenges reflect the complexity of the human psyche and the dynamic nature of emotions and personality traits. There are several key open issues.
Effective collaboration between psychologists, computer scientists, neuroscientists, and other experts is essential. Bridging the gap between these disciplines remains a challenge, as each discipline provides unique insights into the study of emotions and personality [1][2]. Existing tools may not accurately capture emotions and personality traits across diverse cultural contexts. The development of culturally sensitive assessment tools and the accounting for cultural variations remain open issues [3].
Real-time emotion detection and personality assessment are crucial for applications such as mental health support and human-computer interaction. The development of tools that can provide timely and accurate assessments in dynamic environments is a challenge [86]. The collection of personal data related to emotions and personality raises significant ethical questions. Balancing the need for data with privacy and ethical considerations remains an ongoing issue [87].
Emotions are inherently subjective and influenced by context [88]. The development of tools that can account for individual subjectivity and situational context is a complex challenge. Integrating data from various sources, such as text, facial expressions, physiological signals, and voice, into a coherent assessment of emotions and personality traits is an open issue. Advanced multimodal fusion techniques are needed [89].
Ensuring the validity and generalizability of assessment tools across diverse populations and contexts is an ongoing challenge [90]. Validation methodologies must evolve to encompass the intricacies of emotions and personality. Machine learning models trained on biased data can perpetuate stereotypes and inequalities [91]. Ensuring fairness in emotion and personality assessment tools is a critical issue that requires attention. The development of tools that users find acceptable and easy to use is vital for adoption. The challenge is in designing interfaces and interactions that are user-friendly and nonintrusive [92].
Understanding how emotions and personality traits change over time and in response to interventions or life events is a significant open issue. Longitudinal studies are needed to address this aspect. Creating tools that can provide personalized insights into emotions and personality traits for individuals is an emerging challenge. Personalization requires the integration of diverse data sources and adaptive algorithms. Ensuring that AI-based tools for emotion and personality assessment adhere to ethical principles, such as transparency, accountability, and fairness, is a pressing issue [93] [94].
Extending the use of emotion and personality assessment tools to domains beyond psychology, such as healthcare, education, and marketing, poses open challenges in adapting and validating these tools for new contexts. Combining the outputs of automated tools with human judgment and expertise is a complex issue [95]. The development of hybrid systems that leverage both automated and human assessments is an open area of research. Emotions are dynamic and can change rapidly. Modeling these dynamics and their impact on decision-making and behavior is a challenging research problem [96].
Addressing these open issues in the development of tools for identifying emotions and personality traits will require ongoing collaboration, multidisciplinary approaches, and innovative research across psychology, computer science, and related fields.
The following table summarizes the key open issues in developing tools for identifying emotions and personality traits:
Table 11
Key Open Issues in Developing Tools for Identifying Emotions and Personality Traits
Open Issue
|
Description
|
Effective Collaboration between Disciplines
|
Bridging the gap between psychologists, computer scientists, neuroscientists, and other experts to leverage diverse insights into emotions and personality.
|
Cultural Sensitivity
|
Developing assessment tools that accurately capture emotions and personality traits across diverse cultural contexts.
|
Real-time Assessment
|
Developing tools capable of providing timely and accurate assessments of emotions and personality traits in dynamic environments, such as mental health support and HCI applications.
|
Ethical Considerations
|
Balancing the collection of personal data related to emotions and personality with privacy and ethical considerations.
|
Subjectivity and Context
|
Developing tools that account for individual subjectivity and situational context in assessing emotions and personality traits.
|
Multimodal Fusion
|
Integrating data from various sources (e.g., text, facial expressions, physiological signals) into a coherent assessment of emotions and personality traits.
|
Validity and Generalizability
|
Ensuring the validity and generalizability of assessment tools across diverse populations and contexts.
|
Fairness
|
Addressing biases in machine learning models used for emotion and personality assessment to ensure fairness.
|
User Acceptance
|
Designing user-friendly interfaces and interactions for emotion and personality assessment tools.
|
Longitudinal Studies
|
Conducting longitudinal studies to understand how emotions and personality traits change over time and in response to interventions.
|
Personalization
|
Developing tools that provide personalized insights into emotions and personality traits for individuals.
|
Ethical AI
|
Ensuring AI-based tools adhere to ethical principles such as transparency, accountability, and fairness.
|
Extending Applications
|
Adapting and validating emotion and personality assessment tools for new contexts beyond psychology.
|
Hybrid Systems
|
Developing hybrid systems that combine automated and human assessment for more accurate results.
|
Modeling Dynamics
|
Modeling the dynamic nature of emotions and their impact on decision-making and behavior.
|
Table 11 provides a structured overview of the key open issues, allowing for easy reference and understanding of the challenges in developing tools for identifying emotions and personality traits.
RQ3-What types of contributions have been proposed in this research field?
In the research field of identifying emotions and personality traits using computer-based tools, various types of contributions have been proposed. These contributions encompass a wide range of advancements, innovations, and applications. There are some common types of contributions made in this field.
Researchers have proposed novel algorithms and machine learning models for accurately detecting emotions and assessing personality traits. These advancements have led to more robust and reliable tools for automated analysis [63]. Contributions include the development of effective feature extraction methods for different data sources, such as text, speech, facial expressions, and physiological signals. These methods improve the quality of the input data for analysis [62] [63].
Many contributions have focused on integrating data from multiple sources (e.g., text, audio, video) to provide a holistic understanding of emotions and personality traits. Cross-modal fusion techniques have been proposed for this purpose [8]. Researchers have created and shared large datasets containing emotional and personality data, enabling the development and validation of new tools. These databases have contributed to the advancement of the field [9].
The contributions of this study include rigorous validation studies that assess the accuracy and reliability of emotion and personality assessment tools. Validation ensures that the tools are suitable for various applications [6]. Research has been conducted to investigate the cross-cultural applicability of emotion and personality assessment tools. Understanding cultural variations is essential for creating universally valid tools [7].
Contributions have been made in applying emotion and personality assessment tools to healthcare contexts. These tools help in diagnosing and monitoring mental health conditions and providing personalized interventions [10]. In the field of HCIs, contributions involve the development of user interfaces and systems that adapt based on users' emotional states and personality traits, enhancing user experiences [11].
Researchers have proposed personalized approaches to emotion and personality assessment, tailoring recommendations and interventions based on individual profiles [97]. Ethical contributions address the responsible use of tools for emotion and personality assessment. Ethical guidelines and frameworks help ensure user privacy and data security [98]. Contributions extend to commercial applications, where emotion and personality assessment tools are integrated into marketing, customer service, and product design to enhance user engagement and satisfaction [99].
Tools for identifying emotions and personality traits are applied in education and training settings to tailor instructional content and support personalized learning experiences [71]. This field advances psychological research by providing new methods and tools for studying emotions and personality traits in controlled and real-world settings. Researchers have proposed predictive models that use emotion and personality data to forecast behaviors, such as consumer choices, social interactions, and mental health outcomes [72].
Contributions include fostering collaboration between psychologists, computer scientists, neuroscientists, and other experts, leading to a more holistic understanding of emotions and personality. The development of open-source software and libraries for emotion and personality analysis allows for broader access and collaboration within the research community [69] [73].
These types of contributions collectively contribute to the advancement of the field of identifying emotions and personality traits, enabling its application in various domains and addressing complex challenges. The ongoing collaboration between researchers and practitioners continues to drive innovation in this interdisciplinary field.
Table 12
Types of Contributions in the Research Field of Identifying Emotions and Personality Traits
Type of Contribution
|
Description
|
Reference
|
Novel Algorithms
|
Researchers propose innovative algorithms and models for emotion and personality assessment, enhancing automated analysis..
|
[86]
|
Data Integration
|
Efforts focus on integrating data from various sources to provide a holistic understanding of emotions and personality traits, contributing to comprehensive analysis.
|
[87]
|
Validation Studies
|
Contributions include rigorous validation studies to assess the accuracy and reliability of emotion and personality assessment tools, enhancing their credibility.
|
[88]
|
Cross-Cultural Research
|
Research investigates the cross-cultural applicability of emotion and personality assessment tools, aiming to ensure universal validity and inclusivity.
|
[89]
|
Healthcare Applications
|
Contributions involve the application of emotion and personality assessment tools in healthcare settings, enhancing patient care and mental health treatment.
|
[90]
|
HCI Enhancements
|
Research focuses on developing user interfaces and systems that adapt based on users' emotional states and personality traits, improving user experiences.
|
[91]
|
Personalization
|
Researchers propose personalized approaches to emotion and personality assessment, tailoring interventions based on individual profiles.
|
[92]
|
Ethical Considerations
|
Contributions address ethical considerations in emotion and personality assessment, promoting responsible usage and data security.
|
[93]
|
Commercial Applications
|
Tools for identifying emotions and personality traits are integrated into commercial applications to enhance user engagement and satisfaction.
|
[94]
|
Education Applications
|
Contributions extend to education settings, where tools are used to tailor instructional content and support personalized learning experiences.
|
[95]
|
Table 12 summarizes the various types of contributions made in the research field of identifying emotions and personality traits using computer-based tools.
RQ4- What is the most focal topic of the publication trend in the studies identifying emotions and personality traits during the last five years from the selected studies?
The most focal topic of the publication trend in studies identifying emotions and personality traits during the last five years has been the use of machine learning and artificial intelligence (AI) [63] [64] [65] [67] [68]. This is evidenced by the rapid increase in the number of publications using these methods, as well as the increasing sophistication of the models being developed.
One of the main advantages of using machine learning and AI for emotion and personality recognition is that they can be used to analyze large amounts of data quickly and accurately. This is important because emotions and personality traits can be difficult to identify manually, especially in real time. Another advantage of using machine learning and AI is that they can be used to analyze a wide variety of data types, including text, speech, images, and videos. This allows researchers to develop more comprehensive and accurate models of emotion and personality recognition. Some specific examples of the use
Table 13
Publication trends in studies identifying emotions and personality traits
Most Focal Topic
|
Description
|
Reference
|
Use of Machine Learning and AI
|
Dominant trend in recent research on identifying emotions and personality traits, characterized by the increasing use of machine learning and artificial intelligence (AI) methods.
|
[96]
|
Advantages of Machine Learning and AI
|
Highlights the benefits of using machine learning and AI for emotion and personality recognition, including the ability to analyze large datasets quickly and accurately, and the versatility in handling various data types.
|
[97]
|
Specific Examples of Machine Learning and AI Use
|
Provides examples of machine learning applications in identifying emotions and personality traits, such as facial expression analysis, text and speech analysis, and image and video analysis, with diverse applications across different fields.
|
[98]
|
Potential Applications
|
Discusses the potential applications of machine learning and AI in emotion and personality identification, including human-computer interaction, customer service, mental health diagnosis, and crime prevention, indicating the wide-ranging impact of these technologies.
|
[99]
|
of machine learning and AI for emotion and personality recognition include the following:
Machine learning can be used to identify facial expressions, which are key indicators of emotion. For example, a machine learning model could be trained to identify the facial expressions associated with happiness, sadness, anger, and fear [18] [19] [23]. Machine learning can be used to analyze text for emotional cues, such as the use of certain words or phrases. For example, a machine learning model could be trained to identify tweets that express happiness, sadness, anger, or fear [16] [17] [22]. Machine learning can be used to analyze speech for emotional cues, such as the tone of voice, pitch, and rhythm of speech. For example, a machine learning model could be trained to identify phone calls where the caller is expressing anger or distress [20] [21] [24]. Machine learning can be used to analyze images for emotional cues, such as the expressions of people's faces or the body language of people in a scene. For example, a machine learning model could be trained to identify images that depict happiness, sadness, anger, or fear [18] [19] [23]. Machine learning can be used to analyze videos for emotional cues, such as facial expressions, body language, and tone of voice. For example, a machine learning model could be used to identify videos where people express happiness, sadness, anger, or fear [18] [19] [25]. The use of machine learning and AI for emotion and personality recognition has a wide range of potential applications. For example, it could be used to develop more engaging and personalized human-computer interaction systems, to improve customer service, to develop new diagnostic tools for mental health disorders, and to develop new ways to detect and prevent crime.
Overall, the publication trend in studies identifying emotions and personality traits during the last five years has been marked by the increasing use of machine learning and AI, as well as the development of new and improved methods for data collection, labeling, and ethical considerations.
These findings provide insights into the increasing prominence of machine learning and AI in the study of identifying emotions and personality traits during the last five years, as well as the wide range of potential applications and benefits associated with these technologies.
A. CHANNEL OF PUBLICATION
The "channel of publication" refers to the specific venue or platform through which research findings, papers, articles, and other scholarly works are made publicly available to the academic community and the broader public. The choice of publication channel can significantly impact the visibility, accessibility, and credibility of research. Here, we will discuss the concept of the publication channel in more detail:
Publishing the research findings in reputable academic journals is a common and respected channel for disseminating research in various fields. In this case, the study could be submitted to journals that focus on topics related to artificial intelligence, machine learning, affective computing, psychology, human-computer interaction, or any field closely related to the research. These journals typically require rigorous peer review, ensuring the quality and credibility of the research. Presenting the research at conferences and subsequently publishing the proceedings is common in many academic and technical fields. Depending on the subject matter, conferences related to artificial intelligence, machine learning, emotion analysis, and personality assessment may be suitable. Conference papers offer a platform for presenting research to a specialized audience and can facilitate discussions and feedback from experts in the field. Figure 7 shows the number of systematic mapping studies published from 2019 to 2023.
In-depth studies or comprehensive reviews are often published as books or book chapters. The research could be expanded into a book or contribute a chapter to an edited volume related to the study's focus. This allows for a more extensive exploration of the topic and provides a reference for scholars and practitioners in the field. If the research is part of a doctoral or master's thesis, it can be made accessible through the university's library and online repository. This channel is particularly relevant if the research is conducted within an academic institution. The choice of publication channel should align with the research's objectives, target audience, and the level of detail and rigor needed. Researchers may also consider multiple channels to reach different audiences and maximize the impact of their findings.
B. PUBLICATION TREND
The publication trend for systematic mapping studies of tools to identify emotions and personality traits from 2019 to 2023 shows a steady increase in the number of publications each year. This trend is likely due to a number
of factors. The availability of data on emotions and personality traits is increasing. These data are collected from a variety of sources, including social media, wearable devices, and surveys. The increasing sophistication of machine learning and AI algorithms. These algorithms can be used to analyze large amounts of data to identify patterns and trends that are difficult or impossible to identify manually. There is growing interest in the use of emotion and personality recognition in a variety of applications, such as human-computer interaction, customer service, mental health, and security. Figure 8 shows the number of systematic mapping studies on tools for identifying emotions and personality traits published each year from 2019 to 2023.
As shown in Fig. 8, the number of publications increased from 25 in 2019 to 60 in 2023. This represents a fivefold increase over the five-year period. The following are some of the key findings from the systematic mapping studies published during this period. The majority of the tools surveyed use machine learning and AI algorithms to identify emotions and personality traits [63] [64] [65] [67] [68]. The most common data types used to train machine learning models are text, speech, and images [16] [17] [18] [19] [20]. The most common applications of these tools are human-computer interaction, customer service, mental health, and security. The publication trend for systematic mapping studies on tools to identify emotions and personality traits suggests that this is a rapidly growing field with a wide range of potential applications.