Deep clustering based on deep autoencoder for classifying source samples for aeolian landforms

1. Data Overview
This dataset contains surface sediment samples of four aeolian landforms, namely gobi, mobile sand dunes, dry riverbeds, and Nebkha. Over 40 indicators were analyzed for each sample, and Deep Clustering (DC), a Deep Learning (DL) model, was used to classify the samples of aeolian landforms in the area affected by severe wind erosion in the Qaidam Basin, northwestern China.
2. Data Sources and Collection Methods
All samples were collected at a depth of 0-3 cm below the surface. Each sample consists of 10-20 sub-samples, which were randomly collected from a 10m×10m sampling area. Among them, 14 samples were collected from gobi, 5 from Nebkha, and 10 from mobile dunes. The sediment fraction with a particle size less than 63μm in each sample was tested, and 40 indicators were analyzed using a fully automatic sequential wavelength dispersive X-ray fluorescence spectrometer, including SiO₂, Al₂O₃, Fe₂O₃, MgO, CaO, K₂O, Na₂O, P, S, Cl, Sc, Ti, V, Cr, Mn, Co, Ni, Cu, Zn, As, Br, Rb, Sr, Y, Zr, Nb, Mo, Sn, Sb, Ba, La, Ce, Hf, Ta, W, Pb, Bi, Th, U, F, and Ga.
3. Data Processing and Software
https://github.com/MASaraji/Deep-clustering-based-on-deep-autoencoder-for-classifying-source-samples-for-aeolian-landforms
4. Data Format and Structure
The data structure is in standard Excel format.