TMM-Generated Multilayer SPR Biosensor Dataset for Dengue-Relevant Blood Analyte Detection

Dataset Title: TMM-Generated Multilayer SPR Biosensor Dataset for Dengue-Relevant Blood Analyte Detection Associated Paper: "Machine Learning-Driven Multi-Analyte Performance Prediction and SHAP-Based Explainability of a Multilayer SPR Biosensor for Dengue Detection." [Journal name and Manuscript ID to be added upon submission.] File Contents: File 1 - plasma_ALL.csv (11,424 samples)Complete simulated dataset for the plasma analyte. File 2 - platelet_ALL.csv (11,424 samples)Complete simulated dataset for the platelets analyte. File 3 - hemoglobin_ALL.csv (11,424 samples)Complete simulated dataset for the hemoglobin analyte. All three files share an identical column structure, generated using the Transfer Matrix Method (TMM) at wavelength 633 nm for the ZBLAN/ZnO/Ag/BaTiO3/graphene multilayer structure in the Kretschmann attenuated total reflection configuration. Column Descriptions: config_id - Unique identifier encoding the layer thicknesses for that configurationd2_ZnO_nm - Thickness of the ZnO layer (nm)d3_Ag_nm - Thickness of the Ag layer (nm)d4_BaTiO3_nm - Thickness of the BaTiO3 layer (nm)d5_Gr_nm - Thickness of the graphene layer (nm)theta_res_healthy_deg - SPR resonance angle for the healthy-state refractive index (degrees)theta_res_infected_deg - SPR resonance angle for the dengue-infected-state refractive index (degrees)delta_theta_deg - Shift in resonance angle between healthy and infected states (degrees)Rmin - Minimum reflectance at resonance (dimensionless)R_contrast - Reflectance contrast (dimensionless)Sensitivity - Angular sensitivity of the SPR sensor (deg/RIU)FWHM_deg - Full width at half maximum of the resonance curve (degrees)FOM - Figure of merit (RIU^-1)SNR - Signal-to-noise ratio (dimensionless)QF - Quality factor (dimensionless)label_binary - Binary performance classification label (0/1) derived from sensing performance thresholdslabel_tier - Multi-tier performance classification label (0, 1, or 2) derived from sensing performance thresholds Note on classification labels: label_binary and label_tier are derived categorical labels used for classification-based analysis in the associated manuscript. Please see the manuscript's Methods section for the exact threshold definitions used to assign these labels. Methodology Summary:- Simulation method: Transfer Matrix Method (TMM), Kretschmann attenuated total reflection configuration- Operating wavelength: 633 nm- Layer thickness variation: Four active layers (ZnO, Ag, BaTiO3, graphene) systematically varied within fabrication-feasible ranges- Configurations generated: 11,424 per analyte (3 analytes total: plasma, platelets, hemoglobin)- Feature independence check: Spearman correlation analysis confirmed prior to model training- Validation: Physical consistency further confirmed via finite-element method (FEM) simulations (see manuscript for details) Intended Use:This dataset was used to benchmark twelve machine learning algorithms for simultaneous multi-output regression of six SPR sensing parameters, and to perform SHAP-based explainability analysis identifying the structural determinants of sensor performance. It may be reused for:- Reproducing the reported machine learning results- Benchmarking alternative regression or explainability methods on multilayer SPR sensor data- Extending the analysis to additional analytes or sensor configurations License:This dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0). You are free to share and adapt this data for any purpose, provided appropriate credit is given. Citation:If you use this dataset, please cite:Anik, N.D., Imam, K.T. (2026). Machine Learning-Driven Multi-Analyte Performance Prediction and SHAP-Based Explainability of a Multilayer SPR Biosensor for Dengue Detection. [Journal name, once published]And cite this dataset directly using its Zenodo DOI: https://doi.org/10.5281/zenodo.21272790