The document presents a novel approach to multidimensional data mining aimed at addressing three major weaknesses of existing techniques: the need to rescan entire databases when adding attributes, the loss of important association rules due to granularity issues, and the inability to simultaneously find frequent and infrequent rules. A new data schema and algorithm are proposed, including a structured method for generating association rules that improves efficiency and effectiveness. Experimental results demonstrate the algorithm's scalability and performance superiority over conventional methods.