In recent years, more and more attention has been paid to the daily detection of vital signs. Using impulse radio ultra-wideband (IR-UWB) radars to detect vital signs (respiratory rate and heart rate) has become a research hotspot. In healthcare, a long-term continual monitoring of vital signs is required. But most research works on vital signs detection assume that the subjects always remain stationary, however the random activity of human body is inevitable for a long period of monitoring. This paper proposes a new simple and effective algorithm for vital signs detection using only one IR-UWB radar under random activity of human body, consisting of region of interest detection, respiratory signal extraction, and motion detection. And the vital signs detection is finally realized when the subjects sit at desks and do daily work accompanied with common random body activities, e.g., shaking body, using phones, turning head. The experimental results show that the motion detection can effectively improve the detection accuracy of respiratory rate and heart rate.
Software defect prediction is a popular method for optimizing software testing and improving software quality and reliability. However, software defect datasets usually have quality problems, such as class imbalance and data noise. Oversampling by generating the minority class samples is one of the most well-known methods to improving the quality of datasets; however, it often introduces overfitting noise to datasets. To better improve the quality of these datasets, this paper proposes a method called US-PONR, which uses undersampling to remove duplicate samples from version iterations and then uses oversampling through propensity score matching to reduce class imbalance and noise samples in datasets. The effectiveness of this method was validated in a software prediction experiment that involved 24 versions of software data in 11 projects from PROMISE in noisy environments that varied from 0% to 30% noise level. The experiments showed a significant improvement in the quality of datasets pre-processed by US-PONR in noisy imbalanced datasets, especially the noisiest ones, compared with 12 other advanced dataset processing methods. The experiments also demonstrated that the US-PONR method can effectively identify the label noise samples and remove them.
30 January 2025 in Journal of Systems and Software