Search

Time-series features are the characteristics of data periodically collected over time. The calculation of time-series features helps in understanding the underlying patterns and structure of the data, as well as in visualizing the data. The manual calculation and selection of time-series feature from a large temporal dataset are time-consuming. It requires researchers to consider several signal-processing algorithms and time-series analysis methods to identify and extract meaningful features from the given time-series data. These features are the core of a machine learning-based predictive model and are designed to describe the informative characteristics of the time-series signal. For accurate stress monitoring, it is essential that these features are not only informative but also well-distinguishable and interpretable by the classification models. Recently, a lot of work has been carried out on automating the extraction and selection of times-series features. In this paper, a correlation-based time-series feature selection algorithm is proposed and evaluated on the stress-predict dataset. The algorithm calculates a list of 1578 features of heart rate and respiratory rate signals (combined) using the tsfresh library. These features are then shortlisted to the more specific time-series features using Principal Component Analysis (PCA) and Pearson, Kendall, and Spearman correlation ranking techniques. A comparative study of conventional statistical features (like, mean, standard deviation, median, and mean absolute deviation) versus correlation-based selected features is performed using linear (logistic regression), ensemble (random forest), and clustering (k-nearest neighbours) predictive models. The correlation-based selected features achieved higher classification performance with an accuracy of 98.6% as compared to the conventional statistical feature's 67.4%. The outcome of the proposed study suggests that it is vital to have better analytical features rather than conventional statistical features for accurate stress classification. [ABSTRACT FROM AUTHOR]

Objective: To develop a machine learning algorithm utilizing heart rate variability (HRV) and salivary cortisol to detect the presence of acute stress among pregnant women that may be applied to future clinical research. Methods: ECG signals and salivary cortisol were analyzed from 29 pregnant women as part of a crossover study involving a standardized acute psychological stress exposure and a control non-stress condition. A filter-based features selection method was used to identify the importance of different features [heart rate (HR), time- and frequency-domain HRV parameters and salivary cortisol] for stress assessment and reduce the computational complexity. Five machine learning algorithms were implemented to assess the presence of stress with and without salivary cortisol values. Results: On graphical visualization, an obvious difference in heart rate (HR), HRV parameters and cortisol were evident among 17 participants between the two visits, which helped the stress assessment model to distinguish between stress and non-stress exposures with greater accuracy. Eight participants did not display a clear difference in HR and HRV parameters but displayed a large increase in cortisol following stress compared to the non-stress conditions. The remaining four participants did not demonstrate an obvious difference in any feature. Six out of nine features emerged from the feature selection method: cortisol, three time-domain HRV parameters, and two frequency-domain parameters. Cortisol was the strongest contributing feature, increasing the assessment accuracy by 10.3% on average across all five classifiers. The highest assessment accuracy achieved was 92.3%, and the highest average assessment accuracy was 76.5%. Conclusion: Salivary cortisol contributed a significant increase in accuracy of the assessment model compared to using a range of HRV parameters alone. Our machine learning model demonstrates acceptable accuracy in detection of acute stress among pregnant women when combining salivary cortisol with HR and HRV parameters. [ABSTRACT FROM AUTHOR]

This study looked at novel data sources for cardiovascular risk prediction including detailed lifestyle questionnaire and continuous blood pressure monitoring, using ensemble machine learning algorithms (MLAs). The reference conventional risk score compared against was the Framingham Risk Score (FRS). The outcome variables were low or high risk based on calcium score 0 or calcium score 100 and above. Ensemble MLAs were built based on naive bayes, random forest and support vector classifier for low risk and generalized linear regression, support vector regressor and stochastic gradient descent regressor for high risk categories. MLAs were trained on 600 Southeast Asians aged 21 to 69 years free of cardiovascular disease. All MLAs outperformed the FRS for low and high-risk categories. MLA based on lifestyle questionnaire only achieved AUC of 0.715 (95% CI 0.681, 0.750) and 0.710 (95% CI 0.653, 0.766) for low and high risk respectively. Combining all groups of risk factors (lifestyle survey questionnaires, clinical blood tests, 24-h ambulatory blood pressure and heart rate monitoring) along with feature selection, prediction of low and high CVD risk groups were further enhanced to 0.791 (95% CI 0.759, 0.822) and 0.790 (95% CI 0.745, 0.836). Besides conventional predictors, self-reported physical activity, average daily heart rate, awake blood pressure variability and percentage time in diastolic hypertension were important contributors to CVD risk classification. [ABSTRACT FROM AUTHOR]

This paper proposes a method using statistical local and global features for classifying healthy and murmur heart sound recordings from phonocardiogram signals. Classification requires features extraction step that converts each signal into a sequence of feature vectors composed of static and dynamic energy coefficients computed from overlapped analysis windows. Firstly, we propose, for each heartbeat, to extract local features from the local consecutive regions (1st Sound, Systole, 2nd Sound, Diastole) and global ones from the global region. For each region, the features are the statistical features (mean and standard deviation) computed on the feature vector sequence plus the duration. Secondly, we propose to select the relevant features using filter approach based on mutual information criteria. The extraction and selection methods are validated using K nearest neighbor and Gaussian Mixture Models as classifiers. The classification system were evaluated on a sub-dataset of the public PASCAL heart sounds classifying challenge. Results showed that 12 features selected using the Max-Relevance Min-Redundancy selection strategy were sufficient to explain the two classes with 94.97% classification rate higher than 92.74% state-of-the-art rate. We also showed this selection strategy helped the system to be robust to the testing phase when using automatic segmentation rather than manual segmentation. This work demonstrates that local systolic segment features are the most relevant for murmur/normal classification, regardless of segmentation methods. It also shows that feature selection algorithms have potential to highlight certain relevant regions in signals, which is useful for aided diagnostic systems and basic research. [Display omitted] • Local and global statistical features are extracted from PCG beat signals. • Local systolic segment features have a major impact on normal/murmur classification. • "Filters" strategy algorithms select local features rather than global features. • The importance of systolic region is highlighted for normal/murmur classification. [ABSTRACT FROM AUTHOR]

• A novel hybrid-based model is proposed for imbalanced ECG data classification. • The imbalanced classification task is solved in an iterative manner. • Samples in each class with high uncertainty are selected to form the training subset. • The label of the test sample is obtained by voting on predictions of all iterations. • The proposed model has excellent and stable performance on imbalanced datasets. This paper classifies non-ectopic (N), supraventricular ectopic (S), ventricular ectopic (V), and fusion (F) beats in the MIT-BIH arrhythmia database. The classification encounters serious class imbalance since the number of beats in N (majority class) with sample number above the average per class is heavily outnumbered than that in S, V, and F (minority classes) with sample number below the average per class. To address the class imbalance, a novel model based on active training subset selection and modified broad learning system (MBLS) is proposed. In each iteration, the MBLS trained with the current training subset is used to predict the class label of the test sample and actively select a new training subset for the next iteration. Finally, the class of the test sample is determined by voting on the predictions of all iterations. The experimental results show that our method has excellent performance and outperforms the existing methods. [ABSTRACT FROM AUTHOR]