Obviously, prediction results from Lasso penalty regression algorithm have lower
ARV and RMSE compared with principal component regression model, indicating its
stronger generalization ability and higher accuracy.
5 Conclusion
The research shows that the amount of mine gas emission is influenced by many
factors. Its high-dimensional characteristics easily lead to dimension disaster. In order
to eliminate the collinearity of attributes and avoid the over-fitting of functions, Lasso
algorithm is used to reduce the dimension of variables. Finally, the gas emission is
predicted and analyzed on the public data. The results show that the prediction model
based on Lasso has higher accuracy and better generalization performance than the
principal component multiple regression algorithm, and has a wide application prospect
in machine learning.
Acknowledgements. This article is sponsored by National Science and Technology Major
Project of China (2016ZX05045-007-001).
References
1. Qi QJ, Xia SY (2018) Construction of gas emission forecasting sharing platform based on
sub-source prediction method. J Min Saf Environ Prot 45(02):59–64
2. Li ZY (2018) Analysis of distribution law of coal mine gas emission and prediction of GM
(1,1) prediction. J Coal Mine Mod 2018(02):39–41
3. Hu K, Wang SZ, Han S, Wang S (2017) Prediction of gas emission in mining face based on
TLBO-LOIRE. J Appl Basic Eng Sci 25(05):1048–1056
4. Efron B, Hastie T, Johnstone I (2004) Least angle regression. J Math Stat 32(2):407–499
5. Zou H, Trevor H (2005) Regularization and variable selection via the elastic net. J R Stat Soc
67(2):301–320
Table 4. Comparison of two prediction models
Index
Lasso penalty regression algorithm
Principal component analysis
ARV
0.006
0.012
RMSE
3.40
4.89
172
Q. Chen and L. Huang
ARV and RMSE compared with principal component regression model, indicating its
stronger generalization ability and higher accuracy.
5 Conclusion
The research shows that the amount of mine gas emission is influenced by many
factors. Its high-dimensional characteristics easily lead to dimension disaster. In order
to eliminate the collinearity of attributes and avoid the over-fitting of functions, Lasso
algorithm is used to reduce the dimension of variables. Finally, the gas emission is
predicted and analyzed on the public data. The results show that the prediction model
based on Lasso has higher accuracy and better generalization performance than the
principal component multiple regression algorithm, and has a wide application prospect
in machine learning.
Acknowledgements. This article is sponsored by National Science and Technology Major
Project of China (2016ZX05045-007-001).
References
1. Qi QJ, Xia SY (2018) Construction of gas emission forecasting sharing platform based on
sub-source prediction method. J Min Saf Environ Prot 45(02):59–64
2. Li ZY (2018) Analysis of distribution law of coal mine gas emission and prediction of GM
(1,1) prediction. J Coal Mine Mod 2018(02):39–41
3. Hu K, Wang SZ, Han S, Wang S (2017) Prediction of gas emission in mining face based on
TLBO-LOIRE. J Appl Basic Eng Sci 25(05):1048–1056
4. Efron B, Hastie T, Johnstone I (2004) Least angle regression. J Math Stat 32(2):407–499
5. Zou H, Trevor H (2005) Regularization and variable selection via the elastic net. J R Stat Soc
67(2):301–320
Table 4. Comparison of two prediction models
Index
Lasso penalty regression algorithm
Principal component analysis
ARV
0.006
0.012
RMSE
3.40
4.89
172
Q. Chen and L. Huang