<![CDATA[Internal displacement in Nigeria remains a major humanitarian challenge driven by conflict, insurgency, communal clashes, and climate-related disasters. Existing responses are often reactive, highlighting the need for data-driven approaches to support proactive humanitarian planning. This study, titled Data-Driven Modeling and Predictive Analysis of Internal Migration and Displacement Trends in Nigeria for Enhanced Humanitarian Planning, develops a machine learning framework to analyze and predict displacement trends using data from the International Organization for Migration (IOM) Displacement Tracking Matrix, particularly the Baseline Assessment and Needs Monitoring surveys. The dataset comprised over 114,000 records and 19 variables, including demographic characteristics, site accessibility, household size, and priority needs. Data preprocessing involved missing-value handling, categorical standardization, one-hot encoding, and numerical normalization. The highly imbalanced distribution of displacement causes was also considered, with insurgency accounting for 95% of cases, communal clashes 4%, and natural disasters 1%. Exploratory analysis identified Borno, Adamawa, and Yobe as the most affected states, with insurgency consistently dominating and September 2022 recording the highest displacement peak. Logistic Regression, Random Forest, and XGBoost were evaluated. XGBoost achieved the best performance, obtaining a macro F1-score of 0.96 and balanced accuracy of 0.968. It demonstrated strong classification performance for communal clashes (F1 = 0.99) and natural disasters (F1 = 0.90), while correctly identifying nearly all insurgency cases. These findings demonstrate that displacement patterns can be effectively modeled using socio-economic and geographic indicators. The proposed framework can assist humanitarian agencies in anticipating displacement, optimizing resource allocation, and strengthening early-warning systems to improve crisis preparedness and resilience.]]>
