WITHDRAWN: Nonlinear regression framework for geomagnetic data restoration analysis through machine learning techniques

Abstract

To understand the developmental mechanism of the earth’s magnetic field, the credibility of geomagnetic evidence is important as it provides knowledge to be used for near-surface investigations. This paper proposes a technique for earth’s magnetic fields data restoration centered upon methods of machine learning. The conventional approach to linear interpolation is vulnerable to timely inadequacy and high costs, although there is a substantial change in the proposed process. Three conventional learning models, vector support, random forests, and gradient boosts have been developed in this paper. Also, a repetitive, deep study algorithm, the neural network has been studied to further enhance preparation. To describe hyperplane hide away from the consequences of instruction, the proposed learning models were developed. TThe mentioned regression hyperplane that maps the relationship between the incomplete data and the meaning. The defined hyperplane of regression is a relationship mapping between the incomplete and the preserved mock-up details information. Then the educated models, primarily the hyperplanes, have been used to restore the lost geomagnetic traces can be used for further restoration new field information has been obtained. Lastly, there have been statistical analysis analysis derived, y then the informed models have been reconstructed and recovered for more field data for validation with the lack of geomagnetic traces, especially hyperplanes. The trained models were then used for the reconstruction of lost geomagnetic traces, especially hyperplanes, and can be used to reconstruct new information obtained. Finally, numerical analyses have taken place, the results showed that the performance of our approaches was more effective than the traditional approach, as reconstruction accuracy has been updated, and that the reconstruction accuracy efficiency of the proposed method was increased byincrea increased 20 percent.