A predictive model for damp risk in english housing with explainable AI

Area of Science:

• Building Science
• Environmental Health
• Data Science

Background:

• Damp in residential buildings affects indoor air quality, occupant health, and structural integrity, impacting up to 27% of homes in England.
• Identifying at-risk homes is crucial for timely mitigation and preventing escalation of damp-related issues.

Purpose of the Study:

• To develop and evaluate a predictive model for assessing damp risk in residential buildings.
• To identify key building characteristics and energy efficiency indicators associated with damp prevalence.

Main Methods:

• Utilized 2,073 inspection records and national Energy Performance Certificate (EPC) data from a housing association.
• Employed seven machine learning algorithms, evaluating performance on both balanced and imbalanced datasets.
• Applied SHAP (SHapley Additive exPlanations) analysis for model interpretability and identification of key predictors.

Main Results:

• The best-performing machine learning model achieved an accuracy of 0.636 on balanced data and 0.793 on imbalanced data.
• SHAP analysis identified heating cost, energy consumption, and wall energy efficiency as the strongest predictors of damp.
• Statistical and causal analyses provided insights into potential damp risk factors and mitigation strategies.

Conclusions:

• Machine learning models can effectively support the early identification of homes at high risk of developing damp.
• Findings enable housing managers to prioritize interventions, potentially preventing significant damp issues and associated costs.