Semi-supervised vs. supervised learning for mental health monitoring: A case study on bipolar disorder

Casalino, Gabriella  ; Castellano, Giovanna ; Hryniewicz, Olgierd ; Leite, Daniel ; Opara, Karol ; Radziszewska, Weronika ; Kaczmarek-Majer, Katarzyna

International Journal of Applied Mathematics and Computer Science, Tome 33 (2023) no. 3, pp. 419-428.

Résumé

Acoustic features of speech are promising as objective markers for mental health monitoring. Specialized smartphone apps can gather such acoustic data without disrupting the daily activities of patients. Nonetheless, the psychiatric assessment of the patient’s mental state is typically a sporadic occurrence that takes place every few months. Consequently, only a slight fraction of the acoustic data is labeled and applicable for supervised learning. The majority of the related work on mental health monitoring limits the considerations only to labeled data using a predefined ground-truth period. On the other hand, semi-supervised methods make it possible to utilize the entire dataset, exploiting the regularities in the unlabeled portion of the data to improve the predictive power of a model. To assess the applicability of semi-supervised learning approaches, we discuss selected state-of-the-art semi-supervised classifiers, namely, label spreading, label propagation, a semi-supervised support vector machine, and the self training classifier. We use real-world data obtained from a bipolar disorder patient to compare the performance of the different methods with that of baseline supervised learning methods. The experiment shows that semi-supervised learning algorithms can outperform supervised algorithms in predicting bipolar disorder episodes.

Détail

Keywords: semi supervised learning, mental health monitoring, acoustic features, pattern recognition, AI in medicine
Mots-clés : uczenie częściowo nadzorowane, monitorowanie zdrowia psychicznego, cechy akustyczne, rozpoznawanie wzorca

Bibliographie

[1] [1] Alanazi, A. (2022). Using machine learning for healthcare challenges and opportunities, Informatics in Medicine Unlocked 30: 100924.

[2] [2] Antosik-Wójcińska, A.Z., Dominiak, M., Chojnacka, M., Kaczmarek-Majer, K., Opara, K.R., Radziszewska, W., Olwert, A. and Łukasz Świecicki (2020). Smartphone as a monitoring tool for bipolar disorder: A systematic review including data analysis, machine learning algorithms and predictive modelling, International Journal of Medical Informatics 138: 104131.

[3] [3] Ao, X., Luo, P., Ma, X., Zhuang, F., He, Q., Shi, Z. and Shen, Z. (2014). Combining supervised and unsupervised models via unconstrained probabilistic embedding, Information Sciences 257: 101-114.

[4] [4] Arevian, A.C., Bone, D., Malandrakis, N., Martinez, V.R., Wells, K.B., Miklowitz, D.J. and Narayanan, S. (2020). Clinical state tracking in serious mental illness through computational analysis of speech, PLoS ONE 15(1): e0225695.

[5] [5] Arshad, A., Riaz, S. and Jiao, L. (2019). Semi-supervised deep fuzzy c-mean clustering for imbalanced multi-class classification, IEEE Access 7: 28100-28112.

[6] [6] Basu, S., Banerjee, A. and Mooney, R. (2002). Semi-supervised clustering by seeding, Proceedings of the 19th International Conference on Machine Learning (ICML-2002), Sydney, Australia.

[7] [7] Bennett, K. and Demiriz, A. (1998). Semi-supervised support vector machines, in M. Kearns et al. (Eds), Advances in Neural Information Processing Systems, Vol. 11, MIT Press, Cambridge, pp. 368-374.

[8] [8] Bezdek, J.C. (2013). Pattern Recognition with Fuzzy Objective Function Algorithms, Plenum Press, New York.

[9] [9] Bilenko, M., Basu, S. and Mooney, R.J. (2004). Integrating constraints and metric learning in semi-supervised clustering, Proceedings of the 21st International Conference on Machine Learning, Banff, Canada, p. 11.

[10] [10] Breiman, L., Friedman, J.H., Olshen, R.A. and Stone, C.J. (2017). Classification and Regression Trees, Routledge, New York.

[11] [11] Cai, J., Hao, J., Yang, H., Zhao, X. and Yang, Y. (2023). A review on semi-supervised clustering, Information Sciences 632: 164-200.

[12] [12] Casalino, G., Castellano, G., Galetta, F. and Kaczmarek-Majer, K. (2020). Dynamic incremental semi-supervised fuzzy clustering for bipolar disorder episode prediction, in A. Appice et al. (Eds), Discovery Science, DS 2020, Lecture Notes in Computer Science, Vol. 12323, Springer, Cham, pp. 79-93.

[13] [13] Dominiak, M., Kaczmarek-Majer, K., Antosik-Wojcinska, A.Z., Opara, K.R., Wojnar, M., Olwert, A., Radziszewska, W., Hryniewicz, O., Swiecicki, L. and Mierzejewski, P. (2022). Behavioural data collected from smartphones in the assessment of depressive and manic symptoms for bipolar disorder patients: Prospective observational study, Journal of Medical Internet Research 24(1): e28647.

[14] [14] Espinola, C.W., Gomes, J.C., Pereira, J.M.S. and dos Santos, W.P. (2021). Detection of major depressive disorder using vocal acoustic analysis and machine learning-An exploratory study, Research on Biomedical Engineering 37: 53-64.

[15] [15] Eyben, F., Weninger, F., Gross, F. and Schuller, B. (2013). Recent developments in openSMILE, the Munich open-source multimedia feature extractor, Proceedings of the 21st ACM International Conference on Multimedia, Barcelona, Spain, pp. 835-838.

[16] [16] Faurholt-Jepsen, M., Busk, J., Frost, M., Bardram, J.E., Vinberg, M. and Kessing, L.V. (2019). Objective smartphone data as a potential diagnostic marker of bipolar disorder, Australian & New Zealand Journal of Psychiatry 53(2): 119-128, PMID: 30387368.

[17] [17] Faurholt-Jepsen, M., Vinberg, M., Debel, S., Bardram, J.E. and Kessing, L.V. (2016). Behavioral activities collected through smartphones and the association with illness activity in bipolar disorder, International Journal of Methods in Psychiatric Research 25(4): 309-323.

[18] [18] Gomes, H.M., Grzenda, M., Mello, R., Read, J., Le Nguyen, M.H. and Bifet, A. (2022). A survey on semi-supervised learning for delayed partially labelled data streams, ACM Computing Surveys 55(4): 1-42.

[19] [19] González-Almagro, G., Peralta, D., De Poorter, E., Cano, J.-R. and García, S. (2023). Semi-supervised constrained clustering: An in-depth overview, ranked taxonomy and future research directions, arXiv: 2303.00522.

[20] [20] Grande, I., Berk, M., Birmaher, B. and Vieta, E. (2016). Bipolar disorder, The Lancet 387(10027): 1561-1572.

[21] [21] Grünerbl, A., Muaremi, A. and Osmani, V. (2015). Smartphone-based recognition of states and state changes in bipolar disorder patients, IEEE Journal of Biomedical and Health Informatics 19(1): 140-148.

[22] [22] Hryniewicz, O. and Kaczmarek-Majer, K. (2021). Possibilistic aggregation of inhomogeneous streams of data, 2021 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Luxembourg, pp. 1-6, DOI: 10.1109/FUZZ45933.2021.9494583.

[23] [23] Kaczmarek-Majer, K., Casalino, G., Castellano, G., Dominiak, M., Hryniewicz, O., Kamińska, O., Vessio, G. and Díaz-Rodríguez, N. (2022a). Plenary: Explaining black-box models in natural language through fuzzy linguistic summaries, Information Sciences 614: 374-399.

[24] [24] Kaczmarek-Majer, K., Casalino, G., Castellano, G., Hryniewicz, O. and Dominiak, M. (2022b). Explaining smartphone-based acoustic data in bipolar disorder: Semi-supervised fuzzy clustering and relative linguistic summaries, Information Sciences 588: 174-195.

[25] [25] Kaczmarek-Majer, K., Casalino, G., Castellano, G., Leite, D. and Hryniewicz, O. (2022c). Fuzzy linguistic summaries for explaining online semi-supervised learning, 2022 IEEE 11th International Conference on Intelligent Systems, Warsaw, Poland, pp. 1-8.

[26] [26] Kamińska, O., Kaczmarek-Majer, K., Opara, K., Jakuczun, W., Dominiak, M., Antosik-Wójcińska, A., Święcicki, Ł. and Hryniewicz, O. (2019). Self-organizing maps using acoustic features for prediction of state change in bipolar disorder, in M. Marcos et al. (Eds), Artificial Intelligence in Medicine: Knowledge Representation and Transparent and Explainable Systems, Springer, Berlin/Heidelberg, pp. 148-160.

[27] [27] Kamińska, O., Kaczmarek-Majer, K. and Hryniewicz, O. (2020). Acoustic feature selection with fuzzy clustering, self organizing maps and psychiatric assessments, Information Processing and Management of Uncertainty in Knowledge-Based Systems, IPMU 2020, Lisbon, Portugal, pp. 342-355.

[28] [28] Kmita, K., Casalino, G., Castellano, G., Hryniewicz, O. and Kaczmarek-Majer, K. (2022). Confidence path regularization for handling label uncertainty in semi-supervised learning: Use case in bipolar disorder monitoring, 2022 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Padua, Italy, pp. 1-8.

[29] [29] Kusy,M. and Zajdel, R. (2021). A weighted wrapper approach to feature selection, International Journal of Applied Mathematics and Computer Science 31(4): 685-696, DOI: 10.34768/amcs-2021-0047.

[30] [30] Lai, D.T.C. and Garibaldi, J.M. (2011). A comparison of distance-based semi-supervised fuzzy c-means clustering algorithms, 2011 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2011), Taipei, Taiwan, pp. 1580-1586.

[31] [31] Leite, D., Decker, L., Santana, M. and Souza, P. (2020). EGFC: Evolving Gaussian fuzzy classifier from never-ending semi-supervised data streams-With application to power quality disturbance detection and classification, 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Glasgow, UK, pp. 1-9.

[32] [32] Li, K., Cao, Z., Cao, L. and Zhao, R. (2009a). A novel semi-supervised fuzzy c-means clustering method, Chinese Control and Decision Conference, Guilin, China, pp. 3761-3765.

[33] [33] Li, Y.-F., Kwok, J.T. and Zhou, Z.-H. (2009b). Semi-supervised learning using label mean, Proceedings of the 26th Annual International Conference on Machine Learning, Montreal, Canada, pp. 633-640.

[34] [34] Low, D., Bentley, K. and Ghosh, S.K. (2020). Automated assessment of psychiatric disorders using speech: A systematic review, Laryngoscope Investigative Otolaryngology 315(1): 96-116.

[35] [35] Mai, D.S. and Ngo, L.T. (2015). Semi-supervised fuzzy c-means clustering for change detection from multispectral satellite image, IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Istanbul, Turkey, pp. 1-8.

[36] [36] Otsu, N. (1979). A threshold selection method from gray-level histograms, IEEE Transactions on Systems, Man, and Cybernetics 9(1): 62-66.

[37] [37] Panek, D., Skalski, A., Gajda, J. and Tadeusiewicz, R. (2015). Acoustic analysis assessment in speech pathology detection, International Journal of Applied Mathematics and Computer Science 25(3): 631-643, DOI: 10.1515/amcs-2015-0046.

[38] [38] Pedrycz, W. and Waletzky, J. (1997). Fuzzy clustering with partial supervision., IEEE Transactions on Systems, Man and Cybernetics B: Cybernetics 27(5): 787-95.

[39] [39] Ruiz, D. and Finke, J. (2019). Lyapunov-based anomaly detection in preferential attachment networks, International Journal of Applied Mathematics and Computer Science 29(2): 363-373, DOI: 10.2478/amcs-2019-0027.

[40] [40] Vapnik, V. (2006). Estimation of Dependences Based on Empirical Data, Springer Berlin/Heidelberg.

[41] [41] Yarowsky, D. (1995). Unsupervised word sense disambiguation rivaling supervised methods, 33rd Annual Meeting of the Association for Computational Linguistics, Cambridge, USA, pp. 189-196.

[42] [42] Zhou, D., Bousquet, O., Lal, T., Weston, J. and Schölkopf, B. (2003). Learning with local and global consistency, in S. Thrun et al. (Eds), Advances in Neural Information Processing Systems, MIT Press, Cambridge.

[43] [43] Zhu, X. and Ghahramani, Z. (2002). Learning from labeled and unlabeled data with label propagation, Report CMUCALD-02-107, Carnegie Mellon University, Pittsburgh.