Predictive Performance-Explainability Duality for Big Data Analytics-Powered Healthcare

Dear Colleagues,

As Louis-Victor de Broglie proposed the duality principle whereby light is not only a wave but also particle-like, we believe that in order to scale artificial intelligence (AI) with Big Data in healthcare, the duality of rigorous and robust predictive performance evaluation and human-interpretable explainability must be achieved.

Thus, this Special Issue seeks to attract high-quality manuscripts that demonstrate and validate novel contributions to human-interpretable, principled, and reliable predictive performance evaluation with appropriate statistical metrics and explainability, which are key to scale AI-driven applications leveraging Big Data in healthcare sustainably. In particular, this Special Issue builds upon the works of Parisi & Manaog (2023) involving innovative algorithms in machine learning and deep learning in healthcare, the MQAS quantitative assessment scale of papers on AI-driven applications in healthcare, and Chicco & Jurman (2023) on a further validation of the Matthews correlation coefficient (MCC) as a more robust performance evaluation metric for binary classification with imbalanced data, typical of real-life applications in healthcare, than the area under the receiver operating characteristic curve (ROC-AUC).

Therefore, original and thoroughly validated submissions on any of the following topics are welcome and encouraged:

• Novel statistically grounded methodologies to evaluate the predictive performance of machine learning and deep learning for Big Data analytics-powered he
• Big Data classifiers, including (but not limited to) logistic regression, k-nearest neighbour (kNN), ensemble learning approaches, Bayesian classifiers, predictive association classifiers, deep neural networks, classifier chains, besides using the usual metrics such as accuracy and ROC-AUC, which are only applicable in presence of balanced data.
• Big Data regressors, including (but not limited to) kNN with locality-sensitive hashing, ensemble learning approaches, Bayesian regressors, Elastic Net regressors, deep neural networks, regressor chains, besides using the usual metrics such as mean absolute error and Pearson’s correlation coefficient, which are only applicable if larger errors are not more costly and in the presence of (pseudo-)normal data distributions.
• Novel human-interpretable explainability techniques to describe how real-time Big Data analytics-powered predictions are derived on the streamed data fed for inference with respect to the offline or online learning occurred on the training data-related patterns based on:
• Machine-learning-driven algorithms working on distributed databases and with streaming of data, besides the usual SHAP analysis that can be hardly translated into human-explainable terms, thus hindering translational Big Data analytics-powered
• Deep-learning-driven algorithms, besides the usual attention or saliency maps that can be hardly interpreted based on clinicians’ multi-factorial decision-making reasoning and processes, which are instead fully transparent and can be reverse engineered.
• Clustering algorithms and any other unsupervised or self-supervised algorithms, as well as applications that are not related to healthcare, are out of scope of this Special Issue and, thus, any submissions on any of these topics will be rejected.

We are delighted to invite you to submit your high-quality manuscript on any topics mentioned in the summary of our Special Issue entitled “Predictive Performance-Explainability Duality for Big Data Analytics-Powered Healthcare”. The deadline for submitting a full-length paper is 21st June 2024, and we look forward to hearing from you.

Meanwhile, if you have any questions, please do not hesitate to contact us.

References

Chicco, D., Jurman, G. The Matthews correlation coefficient (MCC) should replace the ROC AUC as the standard metric for assessing binary classification. BioData Min. 2023, 16, 1–23.

Parisi, L., RaviChandran, N., Manaog, M.L. A novel hybrid algorithm for aiding prediction of prognosis in patients with hepatitis. Neural Comput. and Appl. 2020, 32, 3839–3852.

Parisi, L., Manaog, M.L. Innovative feature-driven machine learning and deep learning for finance, education, and healthcare. Neural Comput. and Appl. 2023, 35, 11477–11480.

Dr. Luca Parisi
Dr. Mansour Youseffi
Dr. Renfei Ma
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Big Data and Cognitive Computing is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1800 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• predictive performance
• performance evaluation
• machine learning
• deep learning
• explainability
• XAI