Bauxite quality classification by shrinkage methods

Geochemically, bauxite ore may contain some clay minerals, aluminum oxides-hydroxides, and insoluble materials such as quartz and magnetite. The amounts of the geochemical components and their ratios (modules) have critical importance in bauxite quality classification. A classification study was conducted by way of Al2O3/SiO2 module and its corresponding indicators such as spatial coordinates, thickness, and some of the geochemical contributors. The classification was made benefit of two advanced regularization methods such as ridge and the Lasso regression methods. Accuracy, interpretability and simplicity were appraised. Both the algorithms had influence on reducing variance. The smoothing levels of the algorithms also were discussed. The resulting classification provided by the strong estimators can provide a reliable tool for industrial decision making and control.