Abstract
With the development of technology, especially technologies related to artificial intelligence (AI), the fine-dust data acquired by various personal monitoring devices is of great value as training data for predicting future fine-dust concentrations and innovatively alerting people of potential danger. However, most of the fine-dust data obtained from those devices include either missing or abnormal data caused by various factors such as sensor malfunction, transmission errors, or storage errors. This paper presents methods to interpolate the missing data and detect anomalies in PM2.5 time-series data. We validated the performance of our method by comparing ours to well-known existing methods using our personal PM2.5 monitoring data. Our results showed that the proposed interpolation method achieves more than 25% improved results in root mean square error (RMSE) than do most existing methods, and the proposed anomaly detection method achieves fairly accurate results even for the case of the highly capricious fine-dust data. These proposed methods are expected to contribute greatly to improving the reliability of data.
Subject
Fluid Flow and Transfer Processes,Computer Science Applications,Process Chemistry and Technology,General Engineering,Instrumentation,General Materials Science
Cited by
5 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献