Assessing the factors affecting building construction collapse casualty using machine learning techniques: a case of Lagos, Nigeria

Abstract

Building construction collapse in Nigeria has become a subject of international concern in recent times due to numerous lives and properties being wasted yearly. This study presents a brief statistical report of building collapse in Nigeria from 2000–2021, using Lagos State as a case study and conducts a comparative analysis using five supervised machine learning algorithms, namely Robust Linear Model (RLM), Support Vector Machine (SVM), K Nearest Neigbours (KNN), Random Forest (RF) and Decision Tree (DT) for predicting the rate of casualty from building collapse in Lagos Nigeria. Feature importance was performed to determine the most relevant factors that causes building construction collapse casualty. The result shows that the Support Vector Machine (SVM) algorithm has the best forecasting performance among the other algorithms considered. Feature importance analysis, using the SVM model ranked the factors affecting building construction collapse in order of relevance and ‘location’ is considered the most relevant factor contributing to building collapse casualty in Nigeria. Results from this study are important for policy makers and the study recommends that proper onsite geo-technical inspection should be done on site locations before commencement of building constructions in Nigeria.

Keywords:

• Building collapse
• site location
• Lagos
• feature importance
• machine learning

Disclosure statement

The authors declare that there is no competing interest.

Additional information

Funding

The corresponding author acknowledges the State Research Foundation (FAPESP), Sao Paolo, Brazil for support during this research.