Machine Learning Approach for Predicting Coordination Numbers from EXAFS Spectra

This dataset contains the following files:JSON files: Contain the original information for the models. XAS_Energy and XAS_Intensity correspond to the x-axis and y-axis of the XAS spectrum, respectively; group_test and group_test_k correspond to the spectral data in R-space and k-space; average_cn corresponds to the first-shell coordination number of the central atom in the model (since we only study models with integer coordination numbers, data where the coordination number is not an integer should be discarded).CSV files: Contain the raw information for machine learning. The "XX_with_peaks.csv" file contains the x and y coordinates of the XAFS spectrum in R-space, the position and height of the R-space peaks, and the central atom coordination number corresponding to each data point. "FFT_Intensity_xx" describes the y-coordinate value corresponding to the R-space x-coordinate of xx; "peak1_freq" and "peak1_int" describe the x and y coordinates corresponding to the apex of the largest peak in R-space for that spectrum; "coordination_number" corresponds to the central atom coordination number for the spectrum. The "New_XX_k.csv" file contains the x and y coordinates of the XAFS spectrum in k-space and the central atom coordination number corresponding to each data point. "k_xx" describes the y-coordinate value corresponding to the k-space x-coordinate of xx; "coordination_number" corresponds to the central atom coordination number for the spectrum. "q_xx" describes the data values of the XAFS spectrum after being transformed into q-space. The numbers in each column name have no physical meaning and serve only as placeholders for flattening the 2D q-space data into a 1D array; "coordination_number" corresponds to the central atom coordination number for the spectrum.IPYNB files: These are the code files for machine learning. The "FFT_XX_ML.ipynb" file contains the code for using a neural network to learn from the data in the aforementioned CSV files; the "XX_DT.ipynb" file contains the code for using an ensemble decision tree model to learn from the data in the aforementioned CSV files.