Novel MLR-RF-Based Geospatial Techniques: A Comparison with OK

Geostatistical estimation methods rely on experimental variograms that are mostly erratic, leading to subjective model fitting and assuming normal distribution during conditional simulations. In contrast, Machine Learning Algorithms (MLA) (1) free of such limitations, (2) can incorporate information from multiple sources therefore emerge with increasing interest in real-time resource automation. However, MLAs need be explored for robust learning phenomena, better accuracy, computational efficiency. This paper compares MLAs, i.e., Multiple Linear Regression (MLR) Random Forest (RF), Ordinary Kriging (OK). The techniques were applied the publicly available Walkerlake dataset, while exhaustive Walker Lake dataset was validated. results MLR significant (p < 10 × 10−5), correlation coefficients 0.81 (R-square = 0.65) compared 0.79 0.62) RF OK methods. Additionally, automated (free an intermediary step variogram modelling as OK), produced unbiased estimates, identified key samples representing different zones, had higher