Optimizing Business Intelligence Classification Rule Mining Using Quantum-Inspired Genetic Algorithm

This research introduces a novel approach to enhance knowledge discovery through the construction of classification rules for improved event prediction within the domain of business intelligence. One of the applications of the solution can be the improvement of large-scale advertising campaign forecasting. In order to solve the existing problems of data imbalance and high dimensionality in practical applications, we present a new approach of quantum-inspired genetic algorithm (QIGA). Our methodology employs a two-phase strategy: the first phase utilizes QIGA for fast attribute selection while the second phase enhances the manner in which categorical data is represented to enhance rule interpretability. This combined approach is very efficient in producing accurate and usable classification rules without requiring a lot of data pre-processing. Experimental evaluation on benchmark datasets demonstrates the superiority of our proposed method in terms of precision, reliability, and efficiency when compared to traditional approaches. Thus, the proposed algorithm enhances the state-of-art in knowledge discovery by providing a practical and scalable solution to the problem of handling imbalanced data and reducing the effect of high dimensionality, which allows for improving the accuracy of event prediction. Thus, the findings of this research can be applied to various fields other than advertising including fraud detection, customer churn, and medical diagnosis.