A Survey of Data Mining and Social Network to Study the Distribution of Seismic Scale

碩士 === 嶺東科技大學 === 資訊科技系碩士班 === 105 ===   This study used posts of the related affect for earthquakes in the social network to estimate the seismic scale in regions of Taiwan, and compared with the regional seismic scale reported by the Central Weather Bureau (CWB) in Taiwan. Further we are trying to...

Full description

Bibliographic Details
Main Authors: HSIEH, CHIA-FU, 謝佳富
Other Authors: CHUO, YU-JUNG
Format: Others
Language:zh-TW
Published: 2017
Online Access:http://ndltd.ncl.edu.tw/handle/qgrevp
Description
Summary:碩士 === 嶺東科技大學 === 資訊科技系碩士班 === 105 ===   This study used posts of the related affect for earthquakes in the social network to estimate the seismic scale in regions of Taiwan, and compared with the regional seismic scale reported by the Central Weather Bureau (CWB) in Taiwan. Further we are trying to locate the possible area of the epicenter by using a large amount of information that collected from the social network. This study collected the content of posts for related to earthquakes during 2013 to 2016 from the social network PTT. With the increasing and developed of network, the social media became a platform for people to give vent to their feelings. For example, when an earthquake occurs, a lot of people will log into the PTT Gossiping board to post their feelings about the earthquake such as the scale of shaking or discussion the disasters. This study collects the real-time posts for earthquakes to analyze and derives the possible seismic scale by using the data mining method. The content is going to be classified into different regions by mentioned in posts and then collects the interested semantic words for earthquake shaking. In this study, we developed a method to derive the possible seismic scale by analysis the semantic words in decision tree model. We also compared the result with the reporting of seismic scale from the CWB in Taiwan. The result shows a good agreement with the reporting of CWB for most cases except two cases.