TY - JFULL AU - Shun Hattori PY - 2013/3/ TI - Granularity Analysis for Spatio-Temporal Web Sensors T2 - International Journal of Computer and Information Engineering SP - 255 EP - 264 VL - 7 SN - 1307-6892 UR - https://publications.waset.org/pdf/831 PU - World Academy of Science, Engineering and Technology NX - Open Science Index 74, 2013 N2 - In recent years, many researches to mine the exploding Web world, especially User Generated Content (UGC) such as weblogs, for knowledge about various phenomena and events in the physical world have been done actively, and also Web services with the Web-mined knowledge have begun to be developed for the public. However, there are few detailed investigations on how accurately Web-mined data reflect physical-world data. It must be problematic to idolatrously utilize the Web-mined data in public Web services without ensuring their accuracy sufficiently. Therefore, this paper introduces the simplest Web Sensor and spatiotemporallynormalized Web Sensor to extract spatiotemporal data about a target phenomenon from weblogs searched by keyword(s) representing the target phenomenon, and tries to validate the potential and reliability of the Web-sensed spatiotemporal data by four kinds of granularity analyses of coefficient correlation with temperature, rainfall, snowfall, and earthquake statistics per day by region of Japan Meteorological Agency as physical-world data: spatial granularity (region-s population density), temporal granularity (time period, e.g., per day vs. per week), representation granularity (e.g., “rain" vs. “heavy rain"), and media granularity (weblogs vs. microblogs such as Tweets). ER -