models. ⑦ Calculate the average recognition rate for all prediction results. ⑧ Select
the test set parameter that gets the highest recognition rate. Using cross-validation
methods, good learning parameters can be verified based on the characteristics of the
data itself.
4 Application and Analysis of Model
4.1 Selection of Factors Affecting Gas Emission in Working Face
There are many influencing factors affecting the mine gas emission. In this paper, nine
factors measured within 48 months in the mining face of a coal mine in Huaibei
Province are adopted as original variables, including coal seam depth (X 1 ), gas permeability (X 2 ), coal seam thickness (X 3 ), gas content (X 4 ), CH 4 concentration (X 5 ), air
volume (X 6 ), daily production of coal (X 7 ), coal seam distance (X 8 ), and volatile yield
(X 9 ). Table 1 shows 62 sets of data including the amount of gas emission (Y) and its
influence factors in working face. The first 58 sets of data are used as training sets and
the last 4 sets are used to test the prediction effect.
Table 1. Datasets of gas emission quantity and influencing factors
No. X 1
(m)
X 2
(m/d)
X 3
(m)
X 4
(m
3 /t)
X 5
(%)
X 6
(m
3 /min)
X 7
(t)
X 8
(m)
X 9
(%)
Y
(m
3
/t)
1
228 4.1
1.9 4.99 0.22 589
786 10 32 1.38
2
229 6.1
2.5 6.01 0.2 360
786 9
30 1.32
3
230 4.9
1.5 5.27 0.2 405
421 9
30 2.77
4
235 7.2
1.2 4.1
0.2 417.1
1436 10 31 1.84
5
236 6.1
1.1 2.96 0.2 447.2
1517 8
38 0.85
6
237 7.6
3.3 3.7
0.2 477
1644 10 38 1.81
7
238 7.91 1.9 2.01 0.2 480
1616 10 38 0.86
8
239 8.11 1.8 3.11 0.2 462
1694 10 38 0.79
9
240 7.12 1.9 2.14 0.2 462
1661 7
39 1.52
10 241 9.1
1.9 3.25 0.2 480
1661 8
38 1.83
11 242 6.99 2
2.91 0.2 540
1553 7
38 1
12 243 6.72 1.1 2.16 0.2 720
1570 7
39 1.32
13 244 7.81 1.2 3.19 0.2 660
1500 8
37 1.27
14 245 7.37 1.6 4.1
0.24 648
1619 8
34 1.38
15 246 8.26 1.6 2.31 0.25 598.5
1581 9
38 1.36
16 247 8.46 1
2.46 0.2 452.4
1574 6
37 0.83
… … …
… …
… …
…
… … …
168
Q. Chen and L. Huang
the test set parameter that gets the highest recognition rate. Using cross-validation
methods, good learning parameters can be verified based on the characteristics of the
data itself.
4 Application and Analysis of Model
4.1 Selection of Factors Affecting Gas Emission in Working Face
There are many influencing factors affecting the mine gas emission. In this paper, nine
factors measured within 48 months in the mining face of a coal mine in Huaibei
Province are adopted as original variables, including coal seam depth (X 1 ), gas permeability (X 2 ), coal seam thickness (X 3 ), gas content (X 4 ), CH 4 concentration (X 5 ), air
volume (X 6 ), daily production of coal (X 7 ), coal seam distance (X 8 ), and volatile yield
(X 9 ). Table 1 shows 62 sets of data including the amount of gas emission (Y) and its
influence factors in working face. The first 58 sets of data are used as training sets and
the last 4 sets are used to test the prediction effect.
Table 1. Datasets of gas emission quantity and influencing factors
No. X 1
(m)
X 2
(m/d)
X 3
(m)
X 4
(m
3 /t)
X 5
(%)
X 6
(m
3 /min)
X 7
(t)
X 8
(m)
X 9
(%)
Y
(m
3
/t)
1
228 4.1
1.9 4.99 0.22 589
786 10 32 1.38
2
229 6.1
2.5 6.01 0.2 360
786 9
30 1.32
3
230 4.9
1.5 5.27 0.2 405
421 9
30 2.77
4
235 7.2
1.2 4.1
0.2 417.1
1436 10 31 1.84
5
236 6.1
1.1 2.96 0.2 447.2
1517 8
38 0.85
6
237 7.6
3.3 3.7
0.2 477
1644 10 38 1.81
7
238 7.91 1.9 2.01 0.2 480
1616 10 38 0.86
8
239 8.11 1.8 3.11 0.2 462
1694 10 38 0.79
9
240 7.12 1.9 2.14 0.2 462
1661 7
39 1.52
10 241 9.1
1.9 3.25 0.2 480
1661 8
38 1.83
11 242 6.99 2
2.91 0.2 540
1553 7
38 1
12 243 6.72 1.1 2.16 0.2 720
1570 7
39 1.32
13 244 7.81 1.2 3.19 0.2 660
1500 8
37 1.27
14 245 7.37 1.6 4.1
0.24 648
1619 8
34 1.38
15 246 8.26 1.6 2.31 0.25 598.5
1581 9
38 1.36
16 247 8.46 1
2.46 0.2 452.4
1574 6
37 0.83
… … …
… …
… …
…
… … …
168
Q. Chen and L. Huang