A Unified Temporal Erasable Itemset Mining Approach
Tzung‐Pei Hong, Hao Chang, Shu-Min Li, Yu‐Chuan Tsai · 2021
Erasable-itemset mining is often utilized by factories in production planning to find combinations of materials which could cause an acceptable loss if all items in the combination are not available. However, product databases can change over time: new materials or products may be introduced and out-of-date materials or products eliminated. Traditional erasable-itemset mining algorithms do not account for this. Thus, when mining erasable itemsets, we take such additional time information into account. Various temporal constraints (itemset lifespan definitions) are also discussed in this paper. We propose a general temporal erasable itemset mining approach, which, can successfully mine the desired results under different constraints. The experimental performance about the execution time and memory consumption of the proposed method is also shown.