Predictive Modeling Of Macroeconomic Trend In Nigeria: An Integrated Approach Of Using Correlation, Cluster-Based Similarity, And Artificial Neural Network

This study analyzed economic indicators, to assess patterns in Nigeria's GDP, poverty rate, and other key variables over a multi-year period, using a combination of predictive modeling, correlation analysis, clustering, and dimensional reduction techniques. The predictive analysis revealed a strong alignment between the observed and predicted values for GDP and poverty rates, with minor error margins and accuracy levels close to 100%, underscoring the model’s robustness in forecasting these economic indicators. Correlation analysis highlighted significant positive relationships between GDP and other variables, particularly importation, export, and exchange rates, while inflation and MPR showed moderate associations. Cluster analysis identified three distinct periods with economic similarity: 2014–2018, marked by stable growth; 2019–2022, likely influenced by external shocks or adjustments; and 2023, standing out as unique, potentially due to significant economic changes. The tDistributed Stochastic Neighbor Embedding (t-SNE) further validated these clusters by revealing distinct separations between the identified periods. The combined results offer insights into Nigeria’s economic patterns, revealing periods of stability and adjustment, and highlighting variables that significantly influence GDP and poverty, aiding policymakers in identifying trends and making data-informed decisions.