Electrocardiogram-based Deep Learning and Clinical Risk Factors to Predict Atrial Fibrillation

Background: Artificial intelligence (AI)-enabled analysis of 12-lead electrocardiograms (ECGs) may facilitate efficient estimation of incident atrial fibrillation (AF) risk. However, it remains unclear whether AI provides meaningful and generalizable improvement in predictive accuracy beyond clinical risk factors for AF. Methods: We trained a convolutional neural network ("ECG-AI") to infer 5-year incident AF risk using 12-lead ECGs in patients receiving longitudinal primary care at Massachusetts General Hospital (MGH). We then fit three Cox proportional hazards models, each composed of: a) ECG-AI 5-year AF probability, b) the Cohorts for Heart and Aging in Genomic Epidemiology AF (CHARGE-AF) clinical risk score, and c) terms for both ECG-AI and CHARGE-AF ("CH-AI"). We assessed model performance by calculating discrimination (area under the receiver operating characteristic curve, AUROC) and calibration in an internal test set and two external test sets (Brigham and Women's Hospital and UK Biobank). Models were recalibrated to estimate 2-year AF risk in the UK Biobank given limited available follow-up. We used saliency mapping to identify ECG features most influential on ECG-AI risk predictions and assessed correlation between ECG-AI and CHARGE-AF linear predictors. Results: The training set comprised 45,770 individuals (age 55±17 years, 53% women, 2,171 AF events), and the test sets comprised 83,162 individuals (age 59±13 years, 56% women, 2,424 AF events). AUROC was comparable using CHARGE-AF (MGH 0.802, 95% CI 0.767-0.836; BWH 0.752, 95% CI 0.741-0.763; UK Biobank 0.732, 95% CI 0.704-0.759) and ECG-AI (MGH 0.823, 95% CI 0.790-0.856; BWH 0.747, 95% CI 0.736-0.759; UK Biobank 0.705, 95% CI 0.673-0.737). AUROC was highest using CH-AI: MGH 0.838, 95% CI 0.807-0.869; BWH 0.777, 95% CI 0.766-0.788; UK Biobank 0.746, 95% CI 0.716-0.776). Calibration error was low using ECG-AI (MGH 0.0212; BWH 0.0129; UK Biobank 0.0035) and CH-AI (MGH 0.012; BWH 0.0108; UK Biobank 0.0001). In saliency analyses, the ECG P-wave had the greatest influence on AI model predictions. ECG-AI and CHARGE-AF linear predictors were correlated (Pearson r MGH 0.61, BWH 0.66, UK Biobank 0.41). Conclusions: AI-based analysis of 12-lead ECGs has similar predictive utility to a clinical risk factor model for incident AF and both approaches are complementary. ECG-AI may enable efficient quantification of future AF risk.

A Stepwise Approach for High-Dimensional Gaussian Graphical Models

We present a stepwise approach to estimate high dimensional Gaussian graphical models. We exploit the relation between the partial correlation coefficients and the distribution of the prediction errors, and parametrize the model in terms of the Pearson correlation coefficients between the prediction errors of the nodes’ best linear predictors. We propose a novel stepwise algorithm for detecting pairs of conditionally dependent variables. We compare the proposed algorithm with existing methods including graphical lasso (Glasso), constrained `l1-minimization(CLIME) and equivalent partial correlation (EPC), via simulation studies and real life applications. In our simulation study we consider several model settings and report the results using different performance measures that look at desirable features of the recovered graph.

Iterative multilinear optimization for planar model fitting under geometric constraints

Planes are the core geometric models present everywhere in the three-dimensional real world. There are many examples of manual constructions based on planar patches: facades, corridors, packages, boxes, etc. In these constructions, planar patches must satisfy orthogonal constraints by design (e.g. walls with a ceiling and floor). The hypothesis is that by exploiting orthogonality constraints when possible in the scene, we can perform a reconstruction from a set of points captured by 3D cameras with high accuracy and a low response time. We introduce a method that can iteratively fit a planar model in the presence of noise according to three main steps: a clustering-based unsupervised step that builds pre-clusters from the set of (noisy) points; a linear regression-based supervised step that optimizes a set of planes from the clusters; a reassignment step that challenges the members of the current clusters in a way that minimizes the residuals of the linear predictors. The main contribution is that the method can simultaneously fit different planes in a point cloud providing a good accuracy/speed trade-off even in the presence of noise and outliers, with a smaller processing time compared with previous methods. An extensive experimental study on synthetic data is conducted to compare our method with the most current and representative methods. The quantitative results provide indisputable evidence that our method can generate very accurate models faster than baseline methods. Moreover, two case studies for reconstructing planar-based objects using a Kinect sensor are presented to provide qualitative evidence of the efficiency of our method in real applications.

Predicting Intraday cryptocurrency returns – A Sparse Signals approach

We test for the existence of sparse and short-lived signals in minute-by-minute cryptocurrency returns. Using a large set of linear as well as non linear predictors and a machine learning technique called the LASSO, we generate 1-minute ahead out of sample return forecasts for ten major cryptocurrencies. The forecasts obtained from the LASSO are statistically superior to those generated by the benchmark models. The LASSO based estimation selects predictors that are sparse and quite short lived.

Admissibility of linear predictor in the extended growth curve model

In the present paper, we first give the definition of the extended growth curve model, then according to the definition of admissible linear predictor and some matrix properties, obtain the necessary and sufficient conditions for a linear predictor to be admissible in the classes of homogenous and inhomogeneous linear predictors, respectively.

Supporting a Healthier Takeaway Meal Choice: Creating a Universal Health Rating for Online Takeaway Fast-Food Outlets

Digital food ordering platforms are used by millions across the world and provide easy access to takeaway fast-food that is broadly, though not exclusively, characterised as energy dense and nutrient poor. Outlets are routinely rated for hygiene, but not for their healthiness. Nutritional information is mandatory in pre-packaged foods, with many companies voluntarily using traffic light labels to support making healthier choices. We wanted to identify a feasible universal method to objectively score takeaway fast-food outlets listed on Just Eat that could provide users with an accessible rating that can infer an outlet’s ‘healthiness’. Using a sample of takeaway outlets listed on Just Eat, we obtained four complete assessments by nutrition researchers of each outlet’s healthiness to create a cumulative score that ranged from 4 to 12. We then identified and manually extracted nutritional attributes from each outlet’s digital menu, e.g., number of vegetables that have the potential to be numerated. Using generalized linear modelling we identified which attributes were linear predictors of an outlet’s healthiness assessment from nutritional researchers. The availability of water, salad, and the diversity of vegetables were positively associated with academic researchers’ assessment of an outlet’s healthiness, whereas the availability of chips, desserts, and multiple meal sizes were negatively associated. This study shows promise for the feasibility of an objective measure of healthiness that could be applied to all outlet listings on Just Eat and other digital food outlet aggregation platforms. However, further research is required to assess the metric’s validity, its desirability and value to users, and ultimately its potential influence on food choice behaviour.

Curves Classification by Using a Local Likelihood Function and Its Practical Usefulness for Real Data

We extend the classical approach in supervised classification based on the local likelihood estimation to the functional covariates case. The estimation procedure of the functional parameter (slope parameter) in the linear model when the covariate is of functional kind is investigated. We show, on simulated as well on real data, that classification error rates estimated using test samples, and the estimation procedure by local likelihood seem to lead to better estimators than the classical kernel estimation. In addition, this approach is no longer assuming that the linear predictors have a specific parametric form. However, this approach also has two drawbacks. Indeed, it was more expensive and slower than the kernel regression. Thus, as mentioned earlier, kernels other than the Gaussian kernel can lead to a divergence of the Newton-Raphson algorithm. In contrast, using a Gaussian kernel, 4 to 6 iterations are then sufficient to achieve convergence.

Do Learning Approaches Set the Stage for Emotional Well-Being in College Students?

The research aim of this paper was two-fold: to generate evidence that personality factors are linear predictors of the variable approaches to learning (a relevant cognitive-motivational variable of Educational Psychology); and to show that each type of learning approach differentially predicts positive or negative achievement emotions, in three learning situations: class time, study time, and testing. A total of 658 university students voluntarily completed validated questionnaires referring to these three variables. Using an ex post facto design, we conducted correlational analyses, regression analyses, and multiple structural predictions. The results showed that Conscientiousness is associated with and predicts a Deep Approach to learning, while also predicting positive achievement emotions. By contrast, Neuroticism is associated with and significantly predicts a Surface Approach to learning, as well as negative achievement emotions. There are important psychoeducational implications in the university context, both for prevention and for self-improvement, and for programs that offer psychoeducational guidance.

Inrush Current Detection based on Linear Prediction

The occurrence of inrush currents in power transformers is directly related with the reduction of its lifetime, as well as safe operation of electric power systems. In this paper, a new method for inrush current detection, based on linear prediction, is proposed. Linear predictors of order 2 and four are implemented, and the linear prediction error is employed as main parameter on the detection process. The methods performance is evaluated making use of a database built from simulations on Alternative Transients Program, and the modeled electric system is based on real distribution network. Aiming to evaluate the reliability of the method, Gaussian noise is added to the signals, and a noise sensitivity analysis is performed. The results indicate viability of linear prediction as a tool for fast and robust detection of inrush currents, which performed successfully in situations with SNR of 55 dB.