Nigussol1-markon/Building-Life-Cycle-Carbon-Prediction-Ethiopia

Machine learning framework for predicting building life cycle carbon emissions using XGBoost, SHAP, PDP, and scenario analysis. # Building-Life-Cycle-Carbon-Prediction-Ethiopia Machine learning framework for predicting building life cycle carbon emissions using XGBoost, SHAP, PDP, and scenario analysis. # Building Life Cycle Carbon Emission Prediction and Decision Support Framework ## Overview This repository contains the source code and analysis workflow developed for the thesis: **"Development of an optimized and interpretable XBGoost-based framework for building life cycle carbon emission prediction and sustainable decision support in the Ethiopian construction sector."** The study integrates Life Cycle Assessment (LCA), machine learning, explainable artificial intelligence (XAI), and scenario analysis to predict building life cycle carbon emissions and support sustainable construction decision-making. ## Research Objectives The repository supports the following objectives: * Quantification of building life cycle carbon emissions using Life Cycle Assessment (LCA). * Identification of key factors influencing carbon emissions. * Development of an optimized XGBoost-based predictive model. * Model interpretation using SHAP (SHapley Additive Explanations) and Partial Dependence Plots (PDP). * Development of a scenario-based decision-support framework for sustainable construction practices. ## Analysis Workflow Historical Building Data → Data Cleaning and Preprocessing → Life Cycle Carbon Emission Quantification → Feature Selection (RFECV) → Hyperparameter Optimization (Bayesian Optimization) → XGBoost Model Development → Model Validation → SHAP Interpretability Analysis → Partial Dependence Plot (PDP) Analysis → Scenario Analysis → Decision-Support Framework ## Repository Structure ```text ├── data/ │ ├── raw_data/ │ └── processed_data/ │ ├── scripts/ │ ├── preprocessing.py │ ├── lca_calculation.py │ ├── rfecv_feature_selection.py │ ├── bayesian_optimization.py │ ├── xgboost_model.py │ ├── shap_analysis.py │ ├── pdp_analysis.py │ └── scenario_analysis.py │ ├── re …