Frequent Pattern Mining (FPM) from streaming data poses significant challenges due to the need for real-time processing, limited memory availability, and handling the evolving nature of data. Existing algorithms often fall short in adapting to successive transactional updates, leading to reduced accuracy, inefficient memory usage, and poor scalability. To address these challenges, this study proposes a Modified Squirrel Search Algorithm (MSSA) tailored for Successive Transaction-Based Frequent Pattern Mining (ST-FPMDS). The primary objective is to enhance the efficiency and accuracy of frequent pattern discovery in real-time streams by modifying the standard Squirrel Search Algorithm. Key enhancements include dynamic transaction windowing, an adaptive fitness function that evolves with streaming input, and a memory-efficient growth mechanism for frequent pattern sets. Experiments were conducted using real-world streaming foreign exchange dataset. Comparative evaluations were carried out against established algorithms like FP-Growth, Eclat, and PSO-based FPM models using Python and MATLAB-based simulation environments. The results demonstrate that MSSA improves accuracy by 8–14%, reduces processing time by up to 32%, and consumes 21% less memory than traditional methods. The proposed MSSA-based ST-FPMDS framework offers a scalable and adaptive solution for high-throughput data streams
