论文部分内容阅读
以安徽省2000年至2015年发生的139起煤矿事故案例死亡人数为数据样本,对比分析不同周期、不同季度安徽省煤矿企业事故特征。考虑到安全的系统属性,近期预测较远期预测更具有真实性,且不同周期数据的权重值也应有所迥异。据此,以季度为分类点,运用加权线性回归模型建立四个季度的回归模型,然后整合成一个加权线性回归模型组,模拟测度2016—2017年各季度煤矿事故死亡状况,并同一般的时间序列预测进行差异对比。结果表明:第四季度的煤矿事故发生率和死亡人数较其他三个季度高;加权线性回归模型组的显著水平及拟合程度明显高于时间序列。
Based on the data of 139 coal mine accidents in Anhui province from 2000 to 2015, the accident characteristics of coal mines in Anhui Province during different periods and different quarters were analyzed. Taking into account the security of the system attributes, the recent forecast more predictive than the more realistic, and the value of different periodic data should also be very different weights. Based on this, using quarters as a classification point, a regression model of four quarters is established by using a weighted linear regression model and then integrated into a weighted linear regression model group to simulate and measure the deaths of coal mine accidents in each quarter of 2016-2017, and the same time Sequence prediction for differences in comparison. The results show that the coal accident rate and death toll in the fourth quarter were higher than those in the other three quarters. The significant level and the fitting degree of the weighted linear regression model group were significantly higher than the time series.