In this work, we present a weighted l1 norm-based Extreme Learning Machine (ELM), namely, enhanced Regularized ELM (eRELM) for regression problems with outliers contaminated data. The enhancement includes the introduction of a weighting scheme based on l1 norm to a regularization term in the optimization problem. Seismogram is typically recorded with high frequency. Therefore, it requires a large amount of storage. Interpolation allows the seismogram to be stored with the same quality with a smaller number of samples. Our proposed method shows its robustness on interpolation of outlier contaminated seismogram outperforming other ELM based regression techniques by achieving 0.001245 of root mean squared error.
|Title of host publication||Advances in Science, Technology and Innovation|
|Number of pages||3|
|State||Published - 2022|
|Name||Advances in Science, Technology and Innovation|
Bibliographical noteFunding Information:
Acknowledgments This work is supported by the Center for Energy and Geo Processing (CeGP) at King Fahd University of Petroleum and Minerals (KFUPM), Dhahran, Saudi Arabia, under Project GTEC1801.
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
- Extreme learning machine
- L norm
- Seismogram interpolation
ASJC Scopus subject areas
- Environmental Chemistry
- Renewable Energy, Sustainability and the Environment