Intelligent Investment Predictor: An Explainable Machine Learning Framework for Investment Sector Preference Prediction

Abstract

With the changing environment and the occurrence of events around the world such as recession, shifting societalneeds and technology among other factors, it has become difficult for investors to determine the right investment sectors.The global economy is developing rapidly and influenced by environmental changes, economic recessions, developmentof social needs, and new life technologies. Due to these factors, the identification of optimal investment sectors becomeschallenging for investors. Traditional models of forecasting often fail to accommodate this complexity because of limitedgranularity and explainability, which requires more intelligent, explanatory, and robust methods. This research proposesthe Intelligent Investment Advisor, a system leveraging Machine Learning (ML), Deep Learning (DL), and ExplainableAI (XAI) to predict promising investment sectors. The framework combines supervised classification algorithms such asLogistic Regression, Random Forest, Support Vector Machine (SVM), Decision Tree, Naive Bayes, K-Nearest Neighbour(KNN), and Gradient Boosting with ensemble approaches like voting and stacking classifiers. In addition, an optimizedRF-SVM hybrid is implemented to boost predictive capability. The workflow includes data preprocessing, model training,hyperparameter optimization, and ensemble integration.To ensure transparent decision-making, the framework incorporates LIME (Local Interpretable Model-agnosticExplanations), enabling stakeholders to understand the reasoning behind sector predictions. Empirical analysis shows thatensemble classifiers perform better than single models, with the stacked classifier achieving Accuracy: 92.8%, Precision:91.5%, Recall: 93.2%, and F1-score: 92.3%. Compared to conventional models averaging 85% accuracy, the proposedsystem improves performance by 7 to 10 percent.These findings show significant computational improvements in predictive accuracy and statistical robustness, providing areliable decision-support tool for investors, policymakers, and financial institutions.