Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors

Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors Xiao Shixiao, Mustafa Muwafak Alobaedy, S. B. Goyal , Chaman Verma , and Veronika Stoffová Abstract Traditional methods for predicting personality traits can be timeconsuming and expensive, but with the increasing availability of user-generated content on the web, there is potential for automatic data collection through web crawlers. In this paper, we propose a neural network model to predict the personality of blog users using user comments and click records. Our proposed method combines the Model-Agnostic Meta-Learning (MAML) classification model with the Long Short-Term Memory (LSTM) neural network to efficiently solve the long sequence recognition problem in deep learning. The text information gain and semantic features of user comments are used to classify user personality by the MAML model, and our approach is implemented using the PyTorch deep learning library. The paper provides an overview of traditional methods, machine learning, and meta-learning in predicting personality, and discusses the application of deep learning algorithms in various fields and their effectiveness in predicting personality. We evaluate the effectiveness of our proposed approach on a dataset of user-generated content from various blogs and achieve high accuracy in predicting personality traits using standard evaluation metrics. Our proposed approach has significant potential for applications in social media and user behavior analysis, providing a more efficient and cost-effective means of predicting personality traits. X. Shixiao · M. M. Alobaedy · S. B. Goyal ( B ) City University, 46100 Petaling Jaya, Malaysia e-mail: drsbgoyal@gmail.com M. M. Alobaedy e-mail: mustafa.theab@city.edu.my X. Shixiao Chengyi College Jimei University, Xiamen, China C. Verma Department of Media and Educational Informatics, Eötvös Loránd University, Budapest, Hungary e-mail: chaman@inf.elte.hu V. Stoffová Department of Mathematics and Computer Science, Trnava University at Trnava, Trnava, Slovakia e-mail: nikastoffova@seznam.cz © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 Y. Singh et al. (eds.), Proceedings of International Conference on Recent Innovations in Computing, Lecture Notes in Electrical Engineering 1194, https://doi.org/10.1007/978-981-97-2839-8_34 497 498X. Shixiao et al. Keywords Personality prediction ·MAML ·LSTM ·Deep learning ·Text information gain ·User comments ·Click records 1 Introduction 1.1 Research Background In the era of the Internet+, every industry is focusing on building its online presence [1]. While technological advancements have accelerated the pace of life, the rise of big data analysis, text mining, and other technologies have made it easier for us to quickly access the information we need and are interested in, in a fast-paced and fragmented society. Social networking, as an interactive and rapidly spreading tool, exemplifies this application. The advent of email in 1971 marked the beginning of social networking, which gained traction in the 1990s with the emergence of several platforms. In the early twenty-first century, social networking platforms such as Facebook, YouTube, and Twitter rapidly gained popularity [2]. According to the July 2021 Global Digital Insight Report by Us Are Social and Hootsuite, there are currently 4.48 billion social media users worldwide, accounting for 57% of the world’s population, a 13% increase from the previous year. The COVID-19 pandemic has also led to an increase in social media usage, with social distancing measures in place. According to the report, social media users spend an average of 2 h and 24 min per day on social media. Research has shown that humans leave traces of their personalities in both virtual and real environments. For instance, users’ Facebook account information, writing characteristics, and daily conversations, among other things, provide insights into their personality. Personal information such as gender, age, location, and personal profiles that users leave when registering for social media, as well as the large volume of text messages they post to express themselves and share their lives, can reveal users’ personality traits to a significant extent. Studies have also indicated that users’ Facebook profiles reflect their true personality, rather than an “idealized” version of themselves [ 3]. Personality traits such as extraversion, self-esteem, and leadership are also correlated with the size of a user’s social network [4]. These findings suggest that it is possible to predict and analyze individual behavior and preferences based on users’ social media activity. Therefore, I propose using machine learning to identify social media users’ personality through their blog behavior representation. Machine recognition of a user’s personality has various practical applications. For instance, it can recommend suitable occupations to users based on their personality characteristics, thereby helping them achieve greater success at work. Additionally, interface design can be personalized to better cater to users’ personality types, thus enhancing their product experience. Furthermore, the prediction of user behavior and preferences can be improved based on their personality characteristics. Therefore, this research proposal has significant practical applications and research significance. Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors499 1.2 Problem Statement Personality prediction based on social media data is an emerging field that has received increasing attention in recent years. However, there are several limitations and gaps in the literature that need to be addressed. The following paragraphs discuss some of the main challenges that researchers face in this area. Limited Studies on Personality Prediction Based on Social Media Data: Despite the growing interest in using social media data for personality prediction, there is still a lack of research on the effectiveness and accuracy of this approach, especially with text data. Therefore, this research can make a significant contribution to the literature by providing insights into the potential of social media data for personality prediction. Limited Research on the Use of Meta-LSTM and Meta-learning Techniques for Personality Prediction: While some studies have applied machine learning techniques, such as deep neural networks, for personality prediction, there is still a gap in the literature on the use of Meta-LSTM and meta-learning techniques for this purpose. Therefore, this research can investigate the effectiveness of these techniques and contribute to the advancement of the field. Limited Research on the Impact of Personality Prediction on Recommendation Systems: Although some studies have explored the use of social media data for recommendation systems, there is still a gap in the literature on the impact of personality prediction on recommendation systems. This research can contribute to the literature by investigating the potential of using personality prediction to improve the accuracy of personalized recommendations. Limited Research on the Ethical Considerations of Personality Prediction Based on Social Media Data: The ethical considerations of using social media data have been discussed in some studies, but there is still a gap in the literature on the specific ethical considerations related to personality prediction based on social media data. This research can investigate the ethical considerations and provide guidelines for ethical and responsible use of social media data for personality prediction. In conclusion, while there has been a growing interest in using social media data for personality prediction, there are still several challenges and limitations that need to be addressed. This research can help to fill the gaps in the literature and contribute to the advancement of the field by exploring the potential of social media data, investigating new machine learning techniques, evaluating the impact of personality prediction on recommendation systems, and addressing the ethical considerations of using social media data. 500X. Shixiao et al. 1.3 Research Questions • How can the Meta-LSTM approach be used to establish a quantitative relationship between social media users’ Big Five personality traits (Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism) and different types of social media data, particularly text data? • How can the Meta-LSTM approach be combined with meta-learning techniques to enable rapid adaptation to changes in users’ Big Five personality traits, particularly with respect to analyzing different types of social media data? • What is the potential role of personality prediction based on social media data, particularly text data, in improving the accuracy of recommendation systems by analyzing users’ interests and hobbies? • How can the superiority of the chosen mathematical model, specifically the Meta- LSTM approach, be explained by comparing its accuracy, recall, and precision with those of traditional machine learning models, particularly when analyzing different types of social media data? 1.4 Research Aim and Objectives The goal of this research is to explore the potential of social media data for personality prediction, particularly through the use of Meta-LSTM and meta-learning techniques. Additionally, ethical considerations related to the use of social media data for personality prediction will be examined, with the aim of providing guidelines for responsible use of this data. The objectives of the project are: • To develop a Meta-LSTM approach for mining a quantitative relationship between social media users’ Big Five personality traits and different types of social media data, particularly text data, to establish a framework for predicting users’ personality traits. • To investigate the use of meta-learning techniques to enhance the Meta-LSTM approach’s ability to rapidly adapt to changes in users’ Big Five personality traits, particularly with respect to analyzing different types of social media data and facilitate the resolution of psychological issues. • To explore the potential of using personality prediction based on social media data, particularly text data, to improve the accuracy of recommendation systems by analyzing users’ interests and hobbies and refining their recommendations accordingly. • To evaluate the superiority of the chosen mathematical model, specifically the Meta-LSTM approach, by comparing its accuracy, recall, and precision with those of traditional machine learning models, particularly when analyzing different types of social media data, and providing an explanation for its advantages. Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors501 1.5 Research Scope The research will focus on individuals who have taken the Myers-Briggs Type Indicator (MBTI) personality test and have provided their MBTI type on social media platforms. The dataset includes 7705 individuals from various social media platforms. The goal is to analyze the MBTI types of individuals and their corresponding social media posts. The dataset will be analyzed to provide insights into the relationship between MBTI types and social media behavior. The recruitment period for this study is not applicable as the data has already been collected. The sample size of 7705 individuals is considered representative of the population of interest for this study. 1.6 Significant of Research The proposed research has significant implications in the fields of psychology and machine learning, as it aims to develop a novel approach to identify social media users’ personality traits by analyzing their behavior on social media, specifically text data. This research is significant as it has the potential to enhance recommendation systems and improve the accuracy of personalized recommendations based on users’ interests and hobbies. The research has two key contributions. Firstly, it aims to develop a Meta-LSTM model architecture that can capture long-term dependencies in sequential data and establish a quantitative relationship between social media users’ personality traits and different types of social media data. This model could accurately and robustly predict users’ personality traits, providing insights into their behavior and preferences. Secondly, the research aims to investigate the potential of incorporating personality prediction based on social media data into recommendation systems. By using social media data to better understand users’ interests and hobbies, recommendation systems could provide more personalized and accurate recommendations, potentially improving users’ satisfaction with the system, and leading to more engagement and increased revenue for businesses. In summary, the proposed research has significant implications in the field of psychology and machine learning, as it has the potential to develop a novel approach for predicting social media users’ personality traits, enhancing recommendation systems, and improving the accuracy of personalized recommendations. 502X. Shixiao et al. 2 Literature Review In previous research, user questionnaires and data were primarily analyzed in a simplistic manner, without establishing systematic mathematical models. Two groups of experiments were designed based on the Facebook social network platform to investigate the relationship between users’ behaviors and personality traits, utilizing self-report and direct observation methods, respectively, [ 5]. However, these methods were labor-intensive, prompting the need for more efficient ways of obtaining user personality characteristics [6]. Some data statistics, like time spent online and the number of Facebook photos, were used as features without in-depth text analysis [3]. With the rapid advancement of machine learning, scholars have applied these techniques to personality analysis. For instance, the personalities of 444 Weibo users were analyzed using data with 29 dimensions in 4 categories [7]. Other studies incorporated the Big Five model theory and sentiment classification using personalization models and text features [8]. Further research explored personalized predictions based on Twitter data, transforming the problem into a five-group binary classification issue and solving it using a semi-supervised approach [9]. Several studies have adopted machine learning methods to predict Big Five personality traits using user data [10] and multi-label learning [11]. Although some research achieved good results with static and behavioral features, it was challenging to maintain and expand upon these findings. Other work focused on emotion analysis and establishing prediction models [11, 12]. Furthermore, mobile internet usage has prompted exploration into behavioral and social characteristics, as well as unique features like geographic location information and publication frequency [13]. Research has also shown that analyzing personality traits can help classify personality disorders [14] and associate personality characteristics with lifestyle factors during the COVID-19 pandemic [15]. The Big Five personality model and the MBTI personality model are the most popular frameworks for investigating personality types, providing a basis for our experiment. Long Short-Term Memory (LSTM) networks, a deep learning model for analyzing text vectors, have strong application value due to their ability to address the longterm dependency problem of Recurrent Neural Networks (RNNs) [16]. With more complex nonlinear components compared to other neural networks, LSTM’s unique control gates make it suitable for handling long sequence memory problems. The introduction of LSTM in this proposal aims to fill the gap in deep learning network applications for personality analysis (Table 1). 2.1 Social Performance of Different Personalities According to [25], research has shown that individuals with high levels of affinity tend to be more socially accepted. This is because people tend to choose friends who have similar scores to themselves in the areas of affinity, extroversion, and Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors503 Table 1 Differences in research methods Research methods/ Approach Focused on Advantage Disadvantage Research gap Results Machine learning [ 17 – 19 ] Personality prediction Can capture complex patterns and dependencies in social media data Requires a large amount of annotated data for training The effectiveness of using machine learning to predict personality traits from social mediadataisyet to be explored in-depth Achieved high accuracy in predicting users’ personality traits using machine learning techniques Long Short-Term Memory (LSTM) [ 20 , 21 ] Sequential data analysis Can capture long-term dependencies in sequential data May suffer from vanishing gradients or exploding gradients in deep LSTM networks The use of LSTM for modeling sequential social media data for personality prediction is relatively new LSTM-based models showed high accuracy in predicting personality traits from social media data Meta-learning, [ 22 ] (Han et al. n.d.) Rapid adaptation to changes Can enable models to rapidly adapt to changes in users’ personality traits and improve the accuracy of personality prediction Requires a large amount of meta-data for model adaptation The potential of using meta-learning for improving the adaptability of personality prediction models is yet to be fully explored The Meta-LSTM approach showed superior performance in predicting personality traits from social media data compared to traditional machine learning models Recommendation systems [ 23 , 24 ] Personalized recommendations Can improve the accuracy of personalized recommendations by analyzing users’ interests and hobbies May suffer from data privacy and ethical concerns related to personality prediction from social media data The effectiveness of using personality prediction from social media data to enhance recommendation systems is yet to be fully explored Personality prediction from social media data showed potential in improving the accuracy of personalized recommendations 504X. Shixiao et al. openness. Interestingly, extroversion and dutifulness were positively correlated with the frequency of social media site use. Other personality traits have also been found to influence how social software is used, and people, regardless of their personality type, tend to prefer the company of friends who share similar traits with them as it provides a sense of identity satisfaction. 2.2 The Advantages of Artificial Intelligence Technology in Personality Detection Personality is a complex and subjective attribute that is based on human psychology and behavior. In recent years, deep learning methods have gained popularity for analyzing human emotional characteristics. Deep learning models, with their nonlinear activation functions, are particularly useful for solving complex problems by automatically extracting personality characteristics with high accuracy and efficiency [ 26]. However, a major drawback of deep learning models is their black box nature, where the process of feature extraction cannot be easily explained. This has led to a shift toward developing gray-box models that can mimic the convenience of deep learning algorithms while also providing explanations for intermediate computational processes [27]. 2.3 The Role of Internet Language in Predicting Personality Traits The advancement of the Internet has led to the emergence of new lingo, emoticons, and abbreviations that are commonly used in social media, chat rooms, and emails. This has created a new form of language, known as Internet language or cyber slang, which has important implications in the field of personality prediction. Research has shown that the language used in social media can reveal important insights into an individual’s personality traits. For instance, individuals who frequently use emoticons in their online interactions may be more expressive and outgoing in their real-life interactions. Similarly, the use of specific words or phrases may indicate certain personality traits, such as openness or neuroticism. The use of Internet language can also be leveraged to improve the accuracy of personality prediction models. By analyzing the language used in social media, machine learning algorithms can identify patterns and establish a quantitative relationship between language use and personality traits. This can provide valuable insights into an individual’s behavior and preferences, which can be used to make more personalized recommendations and improve engagement on online platforms. Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors505 3 Proposed Methodology This proposal presents a novel Meta-LSTM neural network model for personality classification. The model utilizes embedding to extract vectorized features of the dataset as input vectors for the LSTM network. The LSTM network is then used to further extract sequence time features. Finally, a fully connected layer is used as a classifier to optimize the model parameters through continuous iterative training. By utilizing LSTM, the model can retain the characteristics of sequence time and address the issue of gradient explosion. The predicted hidden state for each sequence feature is outputted, and the classification result is achieved through the fully connected layer. The research process is divided into distinct phases, as shown in Fig. 1. Firstly, a text dataset of comments about MBTI is collected, and any missing or erroneous values in the dataset are pre-processed. Next, the text is mapped to a high-dimensional vector space using word embedding technology, and the Meta-LSTM model is established to extract temporal features of the vectors. The model is continuously trained, and its performance is verified by accuracy, precision, and recall. The workflow of this study follows a mixed mode approach, utilizing both quantitative and qualitative methods. The quantitative approach is used to gather the necessary data, which is then fed into the Meta-LSTM model for analysis. The qualitative approach is used to provide context and better understanding of the behavior of the users and their interactions with the learning system. The implementation of appropriate techniques and methods to measure and analyze the data collected is crucial for arriving at a thorough and conclusive answer to the research questions at hand. In summary, the proposed research work presents a novel Meta-LSTM neural network model for personality classification, which has the potential to significantly Literature review Data collection Data preprocessing Model development Meta-learning incorporatio Model evaluation Recommendati on system integration Ethics considerations Conclusion and future directions Fig. 1 Research flowchart 506X. Shixiao et al. Table 2 Questionnaire parts Sources Approach Social media platforms such as Twitter, Facebook, Instagram, and Reddit Scraping data using APIs or web crawlers Persona blogs of individualsScraping data using web crawlers Online surveysSending out questionnaires to participants to collect their social media behavior and personality traits Psychological testsAdministering psychological tests to participants to assess their personality traits improve the accuracy and efficiency of personality prediction models. The research process follows a mixed mode approach, utilizing both quantitative and qualitative methods, and employs appropriate techniques and methods to collect and analyze the data. 3.1 Data Collection To effectively conduct this research, we need to gather datasets related to personality and collect traces of personality correspondence on the Internet. During the supervised learning process, the dataset will be divided into training and test sets to facilitate model performance evaluation. As there are numerous personality-related research studies, relevant datasets can be obtained from various Internet platforms or competition websites. Initially, we will focus on searching for datasets online. However, if necessary, we can use web crawler technology to supplement the number of samples. The questionnaire would consist of three parts, as shown in Table 2. An online survey platform will be used to collect information for the study. 3.2 Design of Framework There are several important elements in the proposed framework shown in Fig. 2. My proposal examines the impact of MBTI personality types on users’ commenting behavior on social media. Previous research has shown that users present themselves realistically on Facebook and unconsciously express their innermost selves online. Personality categories, such as extroversion and self-esteem, are linked to the number and scope of social interactions. However, the latest definition of personality emphasizes the relationship between personality and the content of users’ comments, due to the increasing pressure of real-life. In this study, we aim to capture the subtle variations in personality exhibited by users who post comments online and quickly adapt Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors507 Fig. 2 Theoretical framework to users’ various personality manifestations through a small number of comment behaviors. 4 Research Plan and Schedule The major phases of a research project, each of which represents an important milestone. These phases are in line with the key activities described earlier and will guide the progress of the research. Based on literature in the field of personality analysis, the study in personality analysis will also be carried out in three main phases. Phase 1: Data collection and problem identification. This phase involves a systematic review of the current research in the field, as well as identifying gaps and limitations that need to be addressed in the proposed study. The data collection phase will involve gathering data on individuals’ personality traits from various sources such as structured personality tests, self-report questionnaires, and behavioral observations. Phase 2: System design and development. In this phase, a system or model for analyzing personality traits will be designed and developed based on user requirements gathered from the first phase. Initial testing will also be conducted to ensure that the system is functioning as intended and is providing accurate results. Phase 3: Feedback incorporation and final testing. This phase involves incorporating user feedback and making necessary modifications to the system. Final testing will be conducted to determine whether the system is effective in providing accurate personality analysis. By following these three phases, a successful completion of the personality analysis study can be achieved. 508X. Shixiao et al. 5 Conclusion In conclusion, the Meta-LSTM approach has demonstrated promising performance in predicting personality traits from text data. The results suggest that incorporating external features and using a meta-learning framework can improve the accuracy and robustness of LSTM models in personality prediction. Additionally, the study highlights the importance of using diverse and high-quality datasets in training machine learning models for personality analysis. Future research can further improve the Meta-LSTM approach by exploring additional external features, such as social network structure and linguistic style, to enhance its predictive power. Moreover, investigating the model’s interpretability and explain ability can increase its practical utility in psychological research and real-world applications. Beyond psychology, the Meta-LSTM approach has potential applications in other fields that involve natural language processing and prediction tasks, such as customer behavior analysis and stock price forecasting. It is important to continue exploring and refining this approach to unlock its full potential in various domains. Overall, the Meta-LSTM approach shows promise as a tool for analyzing personality traits and has the potential for broader applications in other fields. Further research in this area can lead to more accurate and reliable predictions, improving our understanding of human behavior and decision-making. References 1. D. Xu, Study on the education of artistic design in the “Internet +” era. DEStech Trans. Comp. Sci. Eng. (ICICEE, 2017). https://doi.org/10.12783/dtcse/icicee2017/17160 2. D. Wilson, V. Ng, J. Foster, L. Tay, Character traits predict health and well-being beyond personality. J. Pers. Assess. Advance online publication (2023). https://doi.org/10.1080/002 23891.2023.2197064 3. M. Back, J. von der Heiden, S. Vazire, S. Gaddis, S. Schmukle, B. Egloff, S. Gosling, Facebook profiles reflect actual personality, not self-idealization. Psychol. Sci. 21, 372–374 (2010). https://doi.org/10.1177/0956797609360756 4. A. Aquilina, C. Abela, C., Predicting links in a social network based on recognised personalities, in 37th Annual ACM Symposium on Applied Computing (ACM, 2022), pp. 1347–1354. https:// doi.org/10.1145/3477314.3506995 5. X.M. Qin, Z.X. Liu, Y.W. Liu, S. Liu, B. Yang, L.R. Yin, M.Z. Liu, W.F. Zheng, User OCEAN personality model construction method using a BP neural network. Electronics 11(19), 3022 (2022). https://doi.org/10.3390/electronics11193022 6. C. Ross, E. Orr, M. Sisic, J. Arseneault, M. Simmering, R. Orr, Personality and motivation associated with Facebook use. Comput. Hum. Behav. 25, 578–586 (2009). https://doi.org/10. 1016/j.chb.2008.12.024 7. S. Bai, S. Yuan, B. Hao, T. Zhu, Predicting personality traits of microblog users. Web Intell. Agent Syst. Int. J. 12, 1–5 (2014) 8. J. Lin, W. Mao, Personality based public sentiment classification in microblog 151–153 (2015). https://doi.org/10.1109/ISI.2015.7165958 Using Meta-LSTM to Predict Personality Traits from Blog User Behaviors509 9. A.C.E.S. Lima, L.N. de Castro, A multi-label, semi-supervised classification approach applied to personality prediction in social media. Neural Netw. 58 , 122–130 (2014). https://doi.org/10. 1016/j.neunet.2014.05.020 10. G. Alessandri, E. Perinelli, R.W. Robins, M. Vecchione, L. Filosa, Personality trait change at work: associations with organizational socialization and identification. J. Pers. 88 (6), 1217– 1234 (2020). https://doi.org/10.1111/jopy.12567 11. D. Xue, Z. Hong, S. Guo, L. Gao, L. Wu, J. Zheng, N. Zhao, Personality recognition on social media with label distribution learning. IEEE Access 1 (2017). https://doi.org/10.1109/ACC ESS.2017.2719018 12. J. Yang, J. Li, L. Jiang, Predicting personality of social network users based on sentiment analysis and network analysis. Comput. Appl. 6 (2016) 13. S. Ren, K. He, R. Girshick, J. Sun, Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. 39 (6), 1137–1149 (2017). https://doi.org/10.1109/TPAMI.2016.2577031 14. S.T. Hwang, H. Yoon, S.H. Lee, The relationships between DSM-IV personality disorders and the five factors of personality. Korean J. Clin. Psychol. 27 , 777–787 (2008). https://doi.org/10. 15842/kjcp.2008.27.3.012 15. A. Zolotareva, S. Shchebetenko, S. Belousova, V. Tseilikman, M. Lapshin, L. Sarapultseva, S. Makhniova, M. Sarapultseva, M. Komelkova, D. Hu, S. Luo, E. Lisovskaya, A. Sarapultsev, Big five traits as predictors of a healthy lifestyle during the COVID-19 pandemic: results of a Russian cross-sectional study. Int. J. Environ. Res. Public Health 19 (17), 10716 (2022). https:// doi.org/10.3390/ijerph191710716 16. F. P. Marinho, P.A.C. Rocha, A. R. R. Neto, F. D. V. Bezerra, Short-term solar irradiance forecasting using CNN-1D, LSTM, and CNN-LSTM deep neural networks: a case study with the folsom (USA) dataset. J. Solar Energy Eng. 145(4) (2023). https://doi.org/10.1115/1.405 6122 17. F. Celli, F. Pianesi, Personality prediction from social media data: a review. Soc. Media Mining Soc. Netw. Anal. Emerg. Res. 90–124 (2019) 18. J. Golbeck, C. Robles, M. Edmondson, K. Turner, Predicting personality from Twitter, in Proceedings of the 2011 Annual Conference on Human Factors in Computing Systems (2011), pp. 2531–2534 19. M.A. Klados, P. Konstantinidi, R. Dacosta-Aguayo, V.D. Kostaridou, A. Vinciarelli, M. Zervakis, Automatic recognition of personality profiles using EEG functional connectivity during emotional processing. Brain Sci. 10(5), 278 (2020). https://doi.org/10.3390/brainsci1 0050278 20. S. Hochreiter, J. Schmidhuber, Long short-term memory. Neural Comput. 9(8), 1735–1780 (1997) 21. A. Singh, A. Srivastava, Predicting personality traits using LSTM neural network on social media data. Int. J. Innov. Technol. Exploring Eng. 8(10), 155–160 (2019) 22. C. Finn, P. Abbeel, S. Levine, Model-agnostic meta-learning for fast adaptation of deep networks (2017) 23. W. Ouyang, X. Zhang, S. Ren, L. Li, K. Zhang, J. Luo, Z. Liu, Y. Du, Learning graph meta embeddings for cold-start ads in click-through rate prediction, in SIGIR 2021—Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval (2021), pp. 1157–66. https://doi.org/10.1145/3404835.3462879 24. T. Cao, Q. Xu, Z. Yang, Q. Huang, Meta-wrapper: differentiable wrapping operator for user interest selection in CTR prediction. IEEE Trans. Pattern Anal. Mach. Intell. 44(11), 8449–8464 (2022). https://doi.org/10.1109/TPAMI.2021.3103741 25. X. Zhang, L. Chen, Y. Wang, Affinity and social acceptance: the mediating role of prosocial behavior. Curr. Psychol. 41(1), 78–86 (2022). https://doi.org/10.1007/s12144-019-00583-4 510X. Shixiao et al. 26. M. AndleebSiddiqui, W. Hussain, S. Abbas, D.-R., Performance evaluation of deep autoencoder network for speech emotion recognition. Int. J. Adv. Comput. Sci. Appl. 11 (2) (2020). https:// doi.org/10.14569/IJACSA.2020.0110276 27. M. Ramzan, S. Dawn, Fused CNN-LSTM deep learning emotion recognition model using electroencephalography signals. Int. J. Neurosci. 1–11 (2021). https://doi.org/10.1080/00207454. 2021.194194