Applying Data Mining to Explore the Circumstances of Fatal Falls in Construction Industry

• Main Authors: Su, Jie-Ying, 蘇倢縈
• Other Authors: Cheng, Ying-Mei
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/75852714913334144410

碩士 === 中國科技大學 === 土木與防災設計研究所 === 100 === According to the preliminary report of task-force inspection for working place which had occupational disasters before by the Labor Investigation Division of New Taipei City government in mid 2012, 100 violations of labor safety regulations were found, in which 48 violations were concerned with the construction industry. Among them, 18 accidents of falling from high places without proper preventive measures in the construction site worry us the most. Therefore, how to prevent f fall accidents is still a vital issue to all of us. This research collected and analyzed major occupational fall disaster cases in Taiwan and Japan.For statistics in Taiwan, the data came from the real cases listed on the web site of the Council of Labor Affair, Executive Yuan(CLA). This research selected 294 real cases of occupational fall accidents from 2003 to 2010 and defined the time factors (month、date)、engineering factors(business、medium、height) and cause factors (direct cause、indirect cause), all together 7 attributes for group analyze. For statistics in Japan, the data were collected from 191 cases provided by the inistry of Health, Labor and Welfare. We defined the engineering factors (business type of disaster、business scale、types of hazardous material (causal agents)、type of engineering、type of disaster、height of falling)、cause factors (man、material、management) , all together 9 attributes for group analyze. This research concluded and labeled all the cases to build the data base as per the 9th amendment (March 2011) of standard category of business by Directorate-General of Budget, Accounting and Statistic, Executive Yuan and the medium category list by major disaster reporting and inspecting regulation, CLA, Executive Yuan, and analyzed them by using the Expectation- Maximization(EM) of WEKA program. To ensure the number of cases in each category is more than 30 cases, the initial category adopted the 6、7、8、9、10 group in Taiwan, and 5、6、7、8 group in Japan.This research found that both Taiwanese and Japanese labors lack safety awareness toward dangerous environments and they are in short of related knowledge of safety and hygiene. This research recommends that the employers should concentrate on discussing and understanding the mechanism of high falls and establish effective measures, in order to achieve the purpose of disaster prevention by using proper personal protective equipment (PPE) and embedding safety awareness among labors.