An Improved FP-Growth Algorithm with Time Decay Factor and Element Attention Weight

Association rule mining (ARM) is an important data processing technique that can reveal potential correlations within data and extract valuable information. Nevertheless, the validity of information poses a challenge in the context of voluminous data. Acknowledging the prevalent focus on recent and specific elements in real-world scenarios. In this work we propose an improved association mining algorithm. First, based on the FP-Growth algorithm, a time decay factor is introduced to assign weights to transactions, reflecting their timeliness. Then, a weighting function is designed for the items of interest, significantly increasing their proportion in the frequent pattern mining outcomes. In addition, the FP-Growth structure is optimized to enhance the running efficiency. Experimental results demonstrate that the proposed algorithm performs well in variant ARM tasks by achieving several times faster running speed at the cost of a small amount of memory.