No abstract

Machine learning has shown utility in detecting patterns within large, unstructured, and complex datasets. One of the promising applications of machine learning is in precision medicine, where disease risk is predicted using patient genetic data. However, creating an accurate prediction model based on genotype data remains challenging due to the so-called “curse of dimensionality” (i.e., extensively larger number of features compared to the number of samples). Therefore, the generalizability of machine learning models benefits from feature selection, which aims to extract only the most “informative” features and remove noisy “non-informative,” irrelevant and redundant features. In this article, we provide a general overview of the different feature selection methods, their advantages, disadvantages, and use cases, focusing on the detection of relevant features (i.e., SNPs) for disease risk prediction.

Sudden steam-driven eruptions strike without warning and are a leading cause of fatalities at touristic volcanoes. Recent deaths following the 2019 Whakaari eruption in New Zealand expose a need for accurate, short-term forecasting. However, current volcano alert systems are heuristic and too slowly updated with human input. Here, we show that a structured machine learning approach can detect eruption precursors in real-time seismic data streamed from Whakaari. We identify four-hour energy bursts that occur hours to days before most eruptions and suggest these indicate charging of the vent hydrothermal system by hot magmatic fluids. We developed a model to issue short-term alerts of elevated eruption likelihood and show that, under cross-validation testing, it could provide advanced warning of an unseen eruption in four out of five instances, including at least four hours warning for the 2019 eruption. This makes a strong case to adopt real-time forecasting models at active volcanoes.

Sensor data quality plays a vital role in Internet of Things (IoT) applications as they are rendered useless if the data quality is bad. This systematic review aims to provide an introduction and guide for researchers who are interested in quality-related issues of physical sensor data. The process and results of the systematic review are presented which aims to answer the following research questions: what are the different types of physical sensor data errors, how to quantify or detect those errors, how to correct them and what domains are the solutions in. Out of 6970 literatures obtained from three databases (ACM Digital Library, IEEE Xplore and ScienceDirect) using the search string refined via topic modelling, 57 publications were selected and examined. Results show that the different types of sensor data errors addressed by those papers are mostly missing data and faults e.g. outliers, bias and drift. The most common solutions for error detection are based on principal component analysis (PCA) and artificial neural network (ANN) which accounts for about 40% of all error detection papers found in the study. Similarly, for fault correction, PCA and ANN are among the most common, along with Bayesian Networks. Missing values on the other hand, are mostly imputed using Association Rule Mining. Other techniques include hybrid solutions that combine several data science methods to detect and correct the errors. Through this systematic review, it is found that the methods proposed to solve physical sensor data errors cannot be directly compared due to the non-uniform evaluation process and the high use of non-publicly available datasets. Bayesian data analysis done on the 57 selected publications also suggests that publications using publicly available datasets for method evaluation have higher citation rates.