MTL-MAR: Raw Features and Labels for Marine Sedimentation Rate Prediction

Datasets(Features and labels) for training machine learning models for the prediction of global mass accumulation rates of sediments in the seafloor. features.zip: contains all the features that are collected for the study. A subset of the features are used for training. The code make_training_features* selects the subset of features labels_mar.zip: contains all the mass accumulation rates and sedimentation rates measurements that are collected for the study. labels_toc.zip: contains all the total organic carbon concentration measurements that are collected for running the multi task learning model. This data has already been published in Parameswaran, et al. 2025. outputs.zip: contains all prediction maps(MAR and TOC( from multi task learning model)) from all the machine learning models. Available in .npy, .mat, .xyz and .nc formats.