In the broad scientific field of time series forecasting, the ARIMA models and their variants have been widely applied for half a century now due to their mathematical simplicity and flexibility in application. However, with the recent advances in the development and efficient deployment of artificial intelligence models and techniques, the view is rapidly changing, with a shift towards machine and deep learning approaches becoming apparent, even without a complete evaluation of the superiority of the new approach over the classic statistical algorithms. Our work constitutes an extensive review of the published scientific literature regarding the comparison of ARIMA and machine learning algorithms applied to time series forecasting problems, as well as the combination of these two approaches in hybrid statistical-AI models in a wide variety of data applications (finance, health, weather, utilities, and network traffic prediction). Our review has shown that the AI algorithms display better prediction performance in most applications, with a few notable exceptions analyzed in our Discussion and Conclusions sections, while the hybrid statistical-AI models steadily outperform their individual parts, utilizing the best algorithmic features of both worlds.

Despite the apparent importance of mental fatigue detection, a reliable application is hindered due to the incomprehensive understanding of the neural mechanisms of mental fatigue. In this paper, we investigated the topological alterations of functional brain networks in the theta band (4 — 7 Hz) of electroencephalography (EEG) data from 40 male subjects undergoing two distinct fatigue-inducing tasks: a low-intensity one-hour simulated driving and a high-demanding half-hour sustained attention task (psychomotor vigilance task, PVT). Behaviorally, subjects demonstrated a robust mental fatigue effect, as reflected by significantly declined performances in cognitive tasks prior and post these two tasks. Furthermore, characteristic path length presented a positive correlation with task duration, which led to a significant increase between the first and last 5 min of both tasks, indicating a fatigue-related disruption in information processing efficiency. However, significantly increased clustering coefficient was revealed only in the driving task, suggesting distinct network reorganizations between the two fatigue-inducing tasks. Moreover, high accuracy (92% for driving; 97% for PVT) was achieved for fatigue classification with apparently different discriminative functional connectivity features. These findings augment our understanding of the complex nature of fatigue-related neural mechanisms and demonstrate the feasibility of using functional connectivity as neural biomarkers for applicable fatigue monitoring.

Efficient classification of mental workload, an important issue in neuroscience, is limited, so far to single task, while cross-task classification remains a challenge. Furthermore, network approaches have emerged as a promising direction for studying the complex organization of the brain, enabling easier interpretation of various mental states. In this paper, using two mental tasks (N-back and mental arithmetic), we present a framework for crossas well as within-task workload discrimination by utilizing multiband electroencephalography (EEG) cortical brain connectivity. In detail, we constructed functional networks in EEG source space in different frequency bands and considering the individual functional connections as classification features, we identified salient feature subsets based on a sequential feature selection algorithm. These connectivity subsets were able to provide accuracy of 87% for cross-task, 88% for N-back task, and 86% for mental arithmetic task. In conclusion, our method achieved to detect a small number of discriminative interactions among brain areas, leading to high accuracy in both within-task and cross-task classifications. In addition, the identified functional connectivity features, the majority of which were detected in frontal areas in theta and beta frequency bands, helped delineate the shared as well as the distinct neural mechanisms of the two mental tasks.

Despite the apparent usefulness of efficient mental workload assessment in various real-world situations, the underlying neural mechanism remains largely unknown and studies of mental workload are limited to well-controlled cognitive tasks using 2D computer screen. In this work, we investigated functional brain network alterations in a simulated flight experiment with three mental workload levels and compared the reorganization pattern between computer screen (2D) and virtual reality (3D) interfaces. We constructed multiband functional networks in EEG source space, which were further assessed in terms of network efficiency and workload classification performances. We found that increased alpha band efficiencies and beta band local efficiency were associated with elevated mental workload levels, while beta band global efficiency exhibited distinct development trends between 2D and 3D interfaces. Furthermore, using a small subset of connectivity features, we achieved a satisfactory multilevel workload classification accuracy in both interfaces (82% for both 2D and 3D). Further inspection of these discriminative connectivity subsets, we found predominant alpha band connectivity features followed by beta and theta band features with different topological patterns between 2D and 3D interfaces. These findings allow for a more comprehensive interpretation of the neural mechanisms of mental workload in relation to realworld assessment.

In the nascent field of neuroergonomics, mental workload assessment is one of the most important issues and has an apparent significance in real-world applications. Although prior research has achieved efficient single-task classification, scatted studies on cross-task mental workload assessment usually result in unsatisfactory performance. Here, we introduce a data-driven analysis framework to overcome the challenges regarding task-independent workload assessment using a fusion of EEG spectral characteristics and unveil the common neural mechanisms underlying mental workload. Specifically, multi-frequency power spectrum and functional connectivity (FC) were estimated for two workload levels in two working-memory tasks performed by 40 healthy participants, subsequently being fed into a machine learning approach to obtain the importance of each feature vector and evaluate classification performance in a cross-task fashion. Our framework achieved a classification accuracy of 0.94 for task-independent mental workload discrimination. Further investigation of the designated features in terms of their spectral and localization properties revealed task-independent common patterns in the neural mechanisms governing workload. In particular, increased workload was associated with elevated frontal delta and theta power but reduced parietal alpha power, whereas FC exhibited complex frequency- and region-dependent alterations. By implication, the employment of the EEG feature fusion emphasized their utility in serving as promising indicators for different workload conditions applications.

Sports injuries pose significant challenges in athlete welfare and team dynamics, particularly in high-intensity sports like soccer. This study used machine learning algorithms to assess non-contact injury risk in professional male soccer players from physiological and mechanical load variables. Twenty-five professional male soccer players with a first-time, non-contact muscle injury were included in this study. Recordings of external load (speed, distance, and acceleration/deceleration data) and internal load (heart rate) were obtained during all training sessions and official matches over a 4-year period. Machine learning model training and evaluation features were calculated for each of nine different metrics for a 28-day period prior to the injury and an equal-length baseline epoch. The acute surge in the values of each workload metric was quantified by the deviation of maximum values from the average, while the variations of cumulative workload over the last four weeks preceding injury were also calculated. Seven features were selected by the model as prominent estimators of injury incidence. Three of the features concerned acute load deviations (number of sprints, training load score-incorporating heart rate and muscle load- and time of heart rate at the 90-100% of maximum). The four cumulative load features were (total distance, high speed and sprint running distance and training load score). The accuracy of the muscle injury risk assessment model was 0.78, with a sensitivity of 0.73 and specificity of 0.85. Our model achieved high performance in injury risk detection using a limited number of training load variables. The inclusion, for the first time, of heart rate related variables in an injury risk assessment model highlights the importance of physiological overload as a contributor to muscle injuries in soccer. By identifying the important parameters, coaches may prevent muscle injuries by controlling surges of training load during training and competition.