Abstract—Emergency incidents forecasting is quite
significant to emergency response in Mega cities. In this paper,
emergency incidents data were collected and data processing
and analysis work were conducted. It is obvious that some rules
exist when data was counted in different time spans (year and
month). With discovered rules regression models were
constructed and one of the current popular data mining
software, Weka, was used to train and test the models. The
results demonstrated that constructed linear regressed
year-model and month-models fit the original data well (the
MARE, mean average relative absolute error is less than 5%).
The year and months' emergency incidents trend can be
predicted based on this model. At the end of this article, some
factors that produce model deviations were discussed from
social activities perspective.
Index Terms—Cross validation, emergency incidents,
forecasting, linear regression.
Nan Gao and Jiting Xu are with Computer Science and Engineering
Department, University of South Carolina, Columbia, SC 29205 USA
(e-mail:gaon@email.sc.edu, xu57@email.sc.edu).
Xueming Shu is with Institute for Public Safety Research, Tsinghua
University, Haidian District, Beijing, 100084 (e-mail:
shuxm@tsinghua.edu.cn).
Cite:Nan Gao, Xueming Shu, Jiting Xu, Biao Wen, Peng Chen, and Peng Wu, "The Study of Quantitative Forecasting Model on City Emergency Incidents," International Journal of Information and Education Technology vol. 3, no. 5, pp. 575-577, 2013.