MetaCost XGBoost Training and Evaluation Dataset with MATBLAB Codes and files for generating proxies

The dataset consists of two curated subsets designed for the classification of alteration types based on geochemical and proxy variables. The traditional dataset (Trad_Train.csv and Trad_Test.csv) is derived directly from the original complete geochemical dataset (alldata.csv) without any missing values. It includes the original geochemical features and serves as a baseline for model training and evaluation. The simulated dataset (Simu_Train.csv and Simu_Test.csv) was generated using custom MATLAB scripts that transform the original geochemical features into proxy variables, simulating conditions based on multiple geostatistical realizations. These proxy values are represented in a Gaussian scale, which may include negative values due to normalization. In the simulated datasets, the target variable Alteration was initially encoded as integers, with the following mapping: 1 = AAA, 2 = IAA, 3 = PHY, 4 = PRO, 5 = PTS, and 6 = UAL. These datasets are intended for evaluating the performance and generalizability of classification models in both raw and proxy-variable scenarios.