A deep learning approach to injury forecasting in NBA basketball

2021 ◽  
pp. 1-12
Author(s):  
Alexander Cohan ◽  
Jake Schuster ◽  
Jose Fernandez
Keyword(s):  
Deep Learning ◽  
Sports Medicine ◽  
Injury Risk ◽  
Small Sample Size ◽  
Model Performance ◽  
Imbalanced Data ◽  
Feature Learning ◽  
Small Sample ◽  
Injury Incidence ◽  
Learning Approaches

Predicting athlete injury risk has been a holy grail in sports medicine with little progress to date due to a variety of factors such as small sample sizes, significantly imbalanced data, and inadequate statistical approaches. Modeling approaches which are not able to account for the multiple interactions across factors can be misleading. We address the small sample size by collecting longitudinal data of NBA player injuries using publicly available data sources and develop a state of the art deep learning model, METIC, to predict future injuries based on past injuries, game activity, and player statistics. We evaluate model performance using metrics appropriate for imbalanced data and find that METIC performs significantly better than other traditional machine learning approaches. METIC uses feature learning to create interactive features which become meaningful in combination with each other. METIC can be used by practitioners and front offices to improve athlete management and reduce injury incidence, potentially saving sports teams millions in revenue due to reduced athlete injuries.

scholarly journals Deep Learning Approaches for Intrusion Detection

2021 ◽  
pp. 50-64
Author(s):  
Azar Abid Salih ◽  
Siddeeq Y. Ameen ◽  
Subhi R. M. Zeebaree ◽  
Mohammed A. M. Sadeeq ◽  
Shakir Fattah Kak ◽  
...  
Keyword(s):  
Deep Learning ◽  
Intrusion Detection ◽  
Computer Security ◽  
Intrusion Detection System ◽  
Detection System ◽  
Feature Learning ◽  
Detailed Comparison ◽  
Learning Approaches ◽  
Network Intrusion ◽  
Research Problems

Recently, computer networks faced a big challenge, which is that various malicious attacks are growing daily. Intrusion detection is one of the leading research problems in network and computer security. This paper investigates and presents Deep Learning (DL) techniques for improving the Intrusion Detection System (IDS). Moreover, it provides a detailed comparison with evaluating performance, deep learning algorithms for detecting attacks, feature learning, and datasets used to identify the advantages of employing in enhancing network intrusion detection.

scholarly journals Improving the Efficacy of Deep-Learning Models for Heart Beat Detection on Heterogeneous Datasets

2021 ◽  
Vol 8 (12) ◽  
pp. 193
Author(s):  
Andrea Bizzego ◽  
Giulio Gabrieli ◽  
Michelle Jin Yee Neoh ◽  
Gianluca Esposito
Keyword(s):  
Deep Learning ◽  
Small Sample Size ◽  
Classification Performance ◽  
Heart Beat ◽  
Small Sample ◽  
Electrocardiogram Signals ◽  
Heterogeneous Datasets ◽  
Beat Detection ◽  
Physiological Markers ◽  
Bioelectric Signals

Deep learning (DL) has greatly contributed to bioelectric signal processing, in particular to extract physiological markers. However, the efficacy and applicability of the results proposed in the literature is often constrained to the population represented by the data used to train the models. In this study, we investigate the issues related to applying a DL model on heterogeneous datasets. In particular, by focusing on heart beat detection from electrocardiogram signals (ECG), we show that the performance of a model trained on data from healthy subjects decreases when applied to patients with cardiac conditions and to signals collected with different devices. We then evaluate the use of transfer learning (TL) to adapt the model to the different datasets. In particular, we show that the classification performance is improved, even with datasets with a small sample size. These results suggest that a greater effort should be made towards the generalizability of DL models applied on bioelectric signals, in particular, by retrieving more representative datasets.

scholarly journals Unravelling Small Sample Size Problems in the Deep Learning World

2020 ◽  
Author(s):  
Rohit Keshari ◽  
Soumyadeep Ghosh ◽  
Saheb Chhabra ◽  
Mayank Vatsa ◽  
Richa Singh
Keyword(s):  
Deep Learning ◽  
Sample Size ◽  
Small Sample Size ◽  
Small Sample

scholarly journals Comparison of Feature Extraction Methods for Physiological Signals for Heat-Based Pain Recognition

Sensors ◽  
2021 ◽  
Vol 21 (14) ◽  
pp. 4838
Author(s):  
Philip Gouverneur ◽  
Frédéric Li ◽  
Wacław M. Adamczyk ◽  
Tibor M. Szikszay ◽  
Kerstin Luedtke ◽  
...  
Keyword(s):  
Machine Learning ◽  
Feature Extraction ◽  
Deep Learning ◽  
Feature Learning ◽  
Recognition System ◽  
Extraction Methods ◽  
Feature Engineering ◽  
Learning Approaches ◽  
Physiological Sensors ◽  
Pain Recognition

While even the most common definition of pain is under debate, pain assessment has remained the same for decades. But the paramount importance of precise pain management for successful healthcare has encouraged initiatives to improve the way pain is assessed. Recent approaches have proposed automatic pain evaluation systems using machine learning models trained with data coming from behavioural or physiological sensors. Although yielding promising results, machine learning studies for sensor-based pain recognition remain scattered and not necessarily easy to compare to each other. In particular, the important process of extracting features is usually optimised towards specific datasets. We thus introduce a comparison of feature extraction methods for pain recognition based on physiological sensors in this paper. In addition, the PainMonit Database (PMDB), a new dataset including both objective and subjective annotations for heat-induced pain in 52 subjects, is introduced. In total, five different approaches including techniques based on feature engineering and feature learning with deep learning are evaluated on the BioVid and PMDB datasets. Our studies highlight the following insights: (1) Simple feature engineering approaches can still compete with deep learning approaches in terms of performance. (2) More complex deep learning architectures do not yield better performance compared to simpler ones. (3) Subjective self-reports by subjects can be used instead of objective temperature-based annotations to build a robust pain recognition system.

scholarly journals Development and Validation of Suspicious Liver Lesion Deep Learning Algorithm for Computed Tomography

2021 ◽  
Author(s):  
Jacob Johnson ◽  
Kaneel Senevirathne ◽  
Lawrence Ngo
Keyword(s):  
Deep Learning ◽  
Learning Algorithm ◽  
Small Sample Size ◽  
Liver Lesion ◽  
The United States ◽  
Detection Algorithm ◽  
Small Sample ◽  
Lesion Detection ◽  
Liver Lesions ◽  
Deep Learning Algorithm

In this work, we report the results of a deep-learning based liver lesion detection algorithm. While several liver lesion segmentation and classification algorithms have been developed, none of the previous work has focused on detecting suspicious liver lesions. Furthermore, their generalizability remains a pitfall due to their small sample size and sample homogeneity. Here, we developed and validated a highly generalizable deep-learning algorithm for detection of suspicious liver lesions. The algorithm was trained and tested on a diverse dataset containing CT exams from over 2,000 hospital sites in the United States. Our final model achieved an AUROC of 0.84 with a specificity of 0.99 while maintaining a sensitivity of 0.33.

scholarly journals RAPID OBJECT DETECTION SYSTEMS, UTILISING DEEP LEARNING AND UNMANNED AERIAL SYSTEMS (UAS) FOR CIVIL ENGINEERING APPLICATIONS

2018 ◽  
Vol XLII-2 ◽  
pp. 391-398 ◽  
Author(s):  
D. Griffiths ◽  
J. Boehm
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Network Architecture ◽  
Model Performance ◽  
Image Understanding ◽  
Detection Algorithm ◽  
Railway Track ◽  
Track Length ◽  
Unmanned Aerial Systems ◽  
Learning Approaches

With deep learning approaches now out-performing traditional image processing techniques for image understanding, this paper accesses the potential of rapid generation of Convolutional Neural Networks (CNNs) for applied engineering purposes. Three CNNs are trained on 275 UAS-derived and freely available online images for object detection of 3m2 segments of railway track. These includes two models based on the Faster RCNN object detection algorithm (Resnet and Incpetion-Resnet) as well as the novel onestage Focal Loss network architecture (Retinanet). Model performance was assessed with respect to three accuracy metrics. The first two consisted of Intersection over Union (IoU) with thresholds 0.5 and 0.1. The last assesses accuracy based on the proportion of track covered by object detection proposals against total track length. In under six hours of training (and two hours of manual labelling) the models detected 91.3 %, 83.1 % and 75.6 % of track in the 500 test images acquired from the UAS survey Retinanet, Resnet and Inception-Resnet respectively. We then discuss the potential for such applications of such systems within the engineering field for a range of scenarios.

scholarly journals Motif-Matching Based Subgraph-Level Attentional Convolutional Network for Graph Classification

2020 ◽  
Vol 34 (04) ◽  
pp. 5387-5394
Author(s):  
Hao Peng ◽  
Jianxin Li ◽  
Qiran Gong ◽  
Yuanxin Ning ◽  
Senzhang Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Social Network ◽  
Spatial Information ◽  
Structural Information ◽  
Feature Learning ◽  
Drug Analysis ◽  
Classification Performance ◽  
Learning Approaches ◽  
Graph Classification ◽  
Convolutional Network

Graph classification is critically important to many real-world applications that are associated with graph data such as chemical drug analysis and social network mining. Traditional methods usually require feature engineering to extract the graph features that can help discriminate the graphs of different classes. Although recently deep learning based graph embedding approaches are proposed to automatically learn graph features, they mostly use a few vertex arrangements extracted from the graph for feature learning, which may lose some structural information. In this work, we present a novel motif-based attentional graph convolution neural network for graph classification, which can learn more discriminative and richer graph features. Specifically, a motif-matching guided subgraph normalization method is developed to better preserve the spatial information. A novel subgraph-level self-attention network is also proposed to capture the different impacts or weights of different subgraphs. Experimental results on both bioinformatics and social network datasets show that the proposed models significantly improve graph classification performance over both traditional graph kernel methods and recent deep learning approaches.

Improvement of Prediction of Noncontact Injury in Elite Australian Footballers With Repeated Exposure to Established High-Risk Workload Scenarios

2018 ◽  
Vol 13 (9) ◽  
pp. 1130-1135 ◽  
Author(s):  
Marcus J. Colby ◽  
Brian Dawson ◽  
Peter Peeling ◽  
Jarryd Heasman ◽  
Brent Rogalski ◽  
...  
Keyword(s):  
High Risk ◽  
Area Under Curve ◽  
Injury Risk ◽  
Perceived Exertion ◽  
Model Performance ◽  
Load Condition ◽  
Rating Of Perceived Exertion ◽  
Injury Incidence ◽  
Maximal Velocity ◽  
Operating Characteristics

Objectives: To assess the effect of multiple high-risk-scenario (HRS) exposures on noncontact injury prediction in elite Australian footballers. Design: Retrospective cohort study. Methods: Sessional workload data (session rating of perceived exertion, global positioning system–derived distance, sprint distance, and maximum velocity) from 1 club (N = 60 players) over 3 seasons were collated; several established HRSs were also defined. Accumulated HRS sessional exposures were calculated retrospectively (previous 1–8 wk). Noncontact injury data were documented. Univariate and multivariate Poisson regression models determined injury incidence rate ratios (IRRs) while accounting for moderating effects (preseason workload volume and playing experience). Model performance was evaluated using receiver operating characteristics (area under curve). Results: Very low (0–8 sessions: IRR = 5.76; 95% confidence interval [CI], 1.69–19.66) and very high (>15 sessions: IRR = 4.70; 95% CI, 1.49–14.87) exposures to >85% of an individual’s maximal velocity over the previous 8 wk were associated with greater injury risk compared with moderate exposures (11–12 sessions) and displayed the best model performance (area under curve = 0.64). A single session corresponding to a very low chronic load condition over the previous week for all workload variables was associated with increased injury risk, with sprint distance (IRR = 3.25; 95% CI, 1.95–5.40) providing the most accurate prediction model (area under curve = 0.63). Conclusions: Minimal exposure to high-velocity efforts (maximum speed exposure and sprint volume) was associated with the greatest injury risk. Being underloaded may be a mediator for noncontact injury in elite Australian football. Preseason workload and playing experience were not moderators of this effect.

scholarly journals Evaluation of Task fMRI Decoding With Deep Learning on a Small Sample Dataset

2021 ◽  
Vol 15 ◽  
Author(s):  
Sunao Yotsutsuji ◽  
Miaomei Lei ◽  
Hiroyuki Akama
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Statistical Power ◽  
Small Sample Size ◽  
Information Leakage ◽  
Small Sample ◽  
Accuracy Evaluation ◽  
Learning Models ◽  
The Impact

Recently, several deep learning methods have been applied to decoding in task-related fMRI, and their advantages have been exploited in a variety of ways. However, this paradigm is sometimes problematic, due to the difficulty of applying deep learning to high-dimensional data and small sample size conditions. The difficulties in gathering a large amount of data to develop predictive machine learning models with multiple layers from fMRI experiments with complicated designs and tasks are well-recognized. Group-level, multi-voxel pattern analysis with small sample sizes results in low statistical power and large accuracy evaluation errors; failure in such instances is ascribed to the individual variability that risks information leakage, a particular issue when dealing with a limited number of subjects. In this study, using a small-size fMRI dataset evaluating bilingual language switch in a property generation task, we evaluated the relative fit of different deep learning models, incorporating moderate split methods to control the amount of information leakage. Our results indicated that using the session shuffle split as the data folding method, along with the multichannel 2D convolutional neural network (M2DCNN) classifier, recorded the best authentic classification accuracy, which outperformed the efficiency of 3D convolutional neural network (3DCNN). In this manuscript, we discuss the tolerability of within-subject or within-session information leakage, of which the impact is generally considered small but complex and essentially unknown; this requires clarification in future studies.

Sign in / Sign up

Export Citation Format

Share Document