Interpolating Lost Spatio-Temporal Data by Web Sensors

We experience various phenomena (e.g., rain, snow, and earthquake) in the physical world, while we carry out various actions (e.g., posting, querying, and e-shopping) in the Web world. Many researches have tried to mine the Web for knowledge about various phenomena in the physical world, and also several Web services using Webmined knowledge have been made available for the public. Meanwhile, the previous papers have introduced various kinds of “Web Sensors” with Temporal Shift, Temporal Propagation, and Geospatial Propagation to sense the Web for knowledge about a targeted physical phenomenon, i.e., to extract its spatiotemporal data sensitively by analyzing big data on the Web (e.g., Web documents, Web query logs, and e-shopping logs), and compared them based on their correlation coefficients with Japan Meteorological Agency’s physically-sensed spatiotemporal statistics to ensure the accuracy of Web-sensed spatiotemporal data sufficiently. As an industrial application of Web Sensors to a problem of the loss or error of physically-sensed spatiotemporal data due to some sort of troubles (e.g., temporary faults of JMA’s observatories), this paper tries to enable Web Sensors to interpolate lost spatiotemporal data of physical statistics by regression analysis. Keywords-Spatiotemporal Data Mining; Big Data Analysis;