Data and codes for X-SSHM

The data and code for manuscript: An explainable spatial interpolation method considering spatial stratified heterogeneity.1.Environments: -MATLAB R2023a-'System Identification Toolbox'-'Mapping Toolbox'-'Statistics and Machine Learning Toolbox'-'Parallel Computing Toolbox'<br>2.Data:-Raw data and the code for the simulated experiment are located in the 'simulated data' folder. -To generate the simulated data, execute 'sysDataGeneration.m'.-The soil organic matter data cannot be shared publicly due to data protection restrictions.<br>3.Description of data files:-'simulateddata.csv', 'simulateddata-train.csv', and 'simulateddata-test.csv': These files contain all samples, training samples and testing samples, respectively.-Each column in the above tables represents the point ID, stratum ID, x-coordinate, y-coordinate, feature 1, feature 2, coefficient 1, coefficient 2, observation value, and the training sample label, in that order.<br>4.Codes-The code for X-SSHM is stored in the 'X-SSHM' folder. -X_SSHM.m: Code for the X_SSHM model, which includs training of intra-stratum and inter-strata learner, fusing of intra-stratum and inter-stratum features, and interpolation of unknown points . Data for explanation can also be obtained from the workspace of MATLAB after interpolation.-generate_st_feature.m: Code for constructing spatial features.-between_samples_gener.m: Code for generating inter-strata samples for each stratum.-gwr.m: Code for Geographically Weighted Regression model (GWR).-bisquare.m: Code for bi-square kernel function for GWR.-guass.m: Code for Guassian kernel function for GWR.-condnum.m: Code for calculating condition number.-fminsearchbnd.m: Code for finding minimal value with bound constraints.-fminsearchcon.m: Extension of fminsearchbnd.m with general inequality constraints.-golden_section.m: Code for golden selection.-gwr_select_bandwidth.m: Code for selecting optimal bandwidth for GWR.-parseagrs.m: Code for parsing name-value pairs.-rsquare.m: Code for calculating RMSE and R square (R2).-startmatlabpool.m and closematlabpool.m: Code for start and close the parallel pool.<br>-To train the model, interpolate and obtain the data for the explanation, execute 'run.m'.<br><br>5.Visualization-The GSHAP data can be visualized use the code in 'visualization' folder.-GSHAP_dependence_graph.m: Code for GSHAP dependence graph.-GSHAP_summary_graph.m: Code for GSHAP summary graph.<br>