A linear prediction based on logarithmic space (theoretically non-linear) is proposed, which is essentially a geometric series. The number of deaths of coal miners decreases year by year in that proportion.
Disaster risk reduction management Module 4: Preparedness, Prevention and Mit...
Mine Death Estimation
1. Prediction of Coal Mine Accidental Deaths for
5 Years Based on 14 Years Data Analysis
Yousuo Joseph Zou1, Jun Steed Huang2, Tong Xu3*, Anne Zou4,
Bruce He5, Xinyi Tao6 and Sam Zhang7
1 University of Guam, USA
2 Southern University of Science and Technology, CHN
3 Jiangsu University, CHN
4 School of Engineering, Vanderbilt University, USA
5 Ryerson University, CAN
6 University of Southern California, USA
7 Queen's University, CAN
October 18-20, 2019, Changsha, China
The 9th International Conference on Computer Engineering and Networks
4. Introduction
• The situation of coal mine safety production is grim, in order to emphasize
the importance of miners' safety, this paper predicted the death toll of China
over the next five years.
• The death toll of 2005-2018 was logarithmically calculated and analyzed by
the linear regression model, then, using the least square method to find its
regression coefficient, and the regression equation is established by using the
obtained regression coefficients.
• The death toll was predicted using the derived regression equation: Over
time, the number of deaths among coal miners has fallen proportionately
from year to year, and the prediction curve conforms to the existing law of
death.
5. Regression Analysis
Regression analysis is a classical method that uses
the relation between two or more quantitative
variables so that a response or outcome variable
can be predicted from the other, or others.
This method has been widely used in various
fields, such as machine learning in computer field,
bioinformation in biological field, etc.
6. Regression Analysis
Classically, a basic regression function has only one predictive variable, and
its regression model is linear. The function can be stated as follows:
where:
is the value of the response variable in the ith trial;
and are regression coefficients;
is a known constant, namely, the value of the predictor variable in the ith trial;
is a random error term with mean E{ }=0 and variance { }= ;
i=1,…,n;
7. Regression Analysis
Since E{ }=0, it follows E{a+Y}=a+E{Y}, so we can get the regression function:
We use Least Square Estimate to find the estimators of regression parameters
and . In this way, we can get the formula as below:
8. Predict Death Toll
From China University of Mining and Technology, we obtain the number of
deaths from various accidents in China's coal mines from 2005 to 2018.
10. Predict Death Toll
find the estimators of regression parameters
we substitute ‘year sequenc’ and ‘death toll in log space’in the estimated
regression function ,
then using =
∑ ( )
∑
and =
∑ ∑
we can get the value of and :
=-0.2377
=9.0516
then we can get the empirical regression model suit for this paper:
12. Error Analysis
In order to evaluate the regression model based on the logarithmic space
established in this paper, we use the following 7 error evaluation indexes. The
names, formulas, and evaluation values of these indexes are shown in the next
page, and the results show that the model is feasible.
13. Error Analysis
Error type Error formula Error value
Total difference of original and theoretically
data
( − ( ))
-877.3451074
Total least square error Q(X) = ∑ ( − ( ))2
1389088.161
Mean value of original data
=
1
14 2130.428571
Total difference of original data and mean
value
( − )
-5.45697E-12
Total variance = ∑ ( − )2
40391957.43
Standard total deviation
6355.466736
Pearson correlation coefficient
∑ ( − ̅)( − )
∑ ( − ̅) ∑ ( − ) 0.993475881
14. Conclusion
--A linear prediction based on logarithmic space (theoretically non-
linear) is proposed, which is essentially a geometric series
change.
--With the passage of time, the number of deaths of coal miners
decreases year by year in proportion.
--Our future work is to use Deep Learning to grasp the fluctuation
pattern that is away from the straight line.
15. ACKNOWLEDGMENT
Thanks go to:
Xin Luo, Prof. Yuhui Shi for discussions and supports!
China Coal Technology & Engineering Group Corp
Changzhou Research Institute verify data.