Cardiovascular diseases (CVDs) kill about 20.5 million people every year. Early prediction can help people to change their lifestyles and to ensure proper medical treatment if necessary. In this research, ten machine learning (ML) classifiers from different categories, such as Bayes, functions, lazy, meta, rules, and trees, were trained for efficient heart disease risk prediction using the full set of attributes of the Cleveland heart dataset and the optimal attribute sets obtained from three attribute evaluators. The performance of the algorithms was appraised using a 10-fold cross-validation testing option. Finally, we performed tuning of the hyperparameter number of nearest neighbors, namely, ‘k’ in the instance-based (IBk) classifier. The sequential minimal optimization (SMO) achieved an accuracy of 85.148% using the full set of attributes and 86.468% was the highest accuracy value using the optimal attribute set obtained from the chi-squared attribute evaluator. Meanwhile, the meta classifier bagging with logistic regression (LR) provided the highest ROC area of 0.91 using both the full and optimal attribute sets obtained from the ReliefF attribute evaluator. Overall, the SMO classifier stood as the best prediction method compared to other techniques, and IBk achieved an 8.25% accuracy improvement by tuning the hyperparameter ‘k’ to 9 with the chi-squared attribute set.
Medical science-related studies have reinforced that the prevalence of coronary heart disease which is associated with the heart and blood vessels has been the most significant cause of health loss and death globally. Recently, data mining and machine learning have been used to detect diseases based on the unique characteristics of a person. However, these techniques have often posed challenges due to the complexity in understanding the objective of the datasets, the existence of too many factors to analyze as well as lack of performance accuracy. This research work is of two-fold effort: firstly, feature extraction and selection. This entails extraction of the principal components, and consequently, the Correlation-based Feature Selection (CFS) method was applied to select the finest principal components of the combined (Cleveland and Statlog) heart dataset. Secondly, by applying datasets to three single and three ensemble classifiers, the best hyperparameters that reflect the pre-eminent predictive outcomes were investigated. The experimental result reveals that hyperparameter optimization has improved the accuracy of all the models. In the comparative studies, the proposed work outperformed related works with an accuracy of 97.91%, and an AUC of 0.996 by employing six optimal principal components selected from the CFS method and optimizing parameters of the Rotation Forest ensemble classifier.
Visual odometry is the process of estimating incremental localization of the camera in 3-dimensional space for autonomous driving. There have been new learning-based methods which do not require camera calibration and are robust to external noise. In this work, a new method that do not require camera calibration called the “windowed pose optimization network” is proposed to estimate the 6 degrees of freedom pose of a monocular camera. The architecture of the proposed network is based on supervised learning-based methods with feature encoder and pose regressor that takes multiple consecutive two grayscale image stacks at each step for training and enforces the composite pose constraints. The KITTI dataset is used to evaluate the performance of the proposed method. The proposed method yielded rotational error of 3.12 deg/100 m, and the training time is 41.32 ms, while inference time is 7.87 ms. Experiments demonstrate the competitive performance of the proposed method to other state-of-the-art related works which shows the novelty of the proposed technique.
Efforts to increase secondary students’ interest to pursue STEM studies to create a scientifically literate society continue to challenge Malaysia’s education s