FuzzyHen: Model Parameters, Complete Rule Base, Derived Outputs, and Reproducibility Materials for Environmental Heat-Risk Scoring in Laying-Hen Houses

Published: 29 September 2026| Version 1 | DOI: 10.17632/cbktg4r84f.1
Contributors:
,
,
,
,

Description

This dataset provides the model parameters, complete fuzzy rule base, derived computational outputs, and reproducibility materials supporting the manuscript “FuzzyHen: An Interpretable Mamdani Framework for Integrated Environmental Heat-Risk Scoring in Reduced-Scale Laying-Hen Houses”. FuzzyHen is an interpretable Mamdani fuzzy inference framework developed to integrate environmental thermal indices into a continuous environmental heat-risk score for laying-hen housing research. The framework uses three redundancy-aware inputs: black globe humidity index (BGHI), temperature-humidity index (THI), and radiant thermal load (RTL). Each input is represented by five linguistic sets, resulting in a complete 125-rule base. Mamdani inference is implemented using minimum antecedent activation, maximum aggregation, and centroid defuzzification over a continuous 0–10 output universe. The deposited materials include a consolidated Excel workbook containing the mathematical model architecture, thermal-index equations, exact input and output membership-function parameters, weight scenarios, the complete 125-rule base, correlation audit, sensitivity analyses, leave-one-day-out robustness results, comparator analyses, computational integrity tests, software environment information, file inventory, data dictionary, and SHA-256 checksums. The accompanying reproducibility package contains the three author-generated source spreadsheets from the controlled reduced-scale experiment, the frozen Python implementation of the FuzzyHen pipeline, the transparent comparator analysis, full record-level derived outputs, machine-readable CSV files, software requirements, and cryptographic checksums. The experimental database comprises 68,792 valid environmental records obtained over 16 common experimental days from three fixed reduced-scale house–roof condition profiles. The dataset supports computational reproducibility and methodological inspection of FuzzyHen. It does not constitute physiological validation, commercial-farm validation, or evidence of generalized causal effects of roofing materials. House identity and roof condition were not independently replicated; therefore, application-level comparisons are restricted to the three observed fixed house–roof profiles.

Files

Institutions

Categories

Computer Science, Artificial Intelligence, Environmental Engineering, Animal Science, Agricultural Engineering

Funders

Licence