IHUOP: Incremental High Utility Occupancy Pattern Discovery

High utility-occupancy pattern mining (HUOPM) provides a nuanced analytical perspective by jointly capturing the utility contribution of patterns and their proportional influence within transactions. When operating in incremental environments, however, maintaining these dual statistics in a coherent and computationally efficient manner becomes substantially more intricate. Existing incremental mining techniques, mainly tailored to frequency- or utility-driven models, are structurally incapable of accommodating the coupled dependencies intrinsic to HUOPM. To operationalize HUOPM in dynamically evolving databases, we propose IHUOP, an incremental HUOPM algorithm based on a compact utility-occupancy list that encapsulates all essential transactional statistics within a unified structure. This design enables the algorithm to assimilate newly arriving transactions without reprocessing historical data, thereby ensuring scalability. Building upon this representation, we devise refined pruning mechanisms based on support estimation and rigorously derived upper bounds of utility occupancy, permitting aggressive yet lossless search-space reduction. Extensive experiments conducted on both real-world and synthetic datasets demonstrate that IHUOP consistently preserves exactness while delivering pronounced improvements in runtime efficiency and memory utilization. These results substantiate the algorithm’s suitability as a robust and scalable solution for dynamic utility-occupancy mining.