Mining condensed representations of frequent itemsets (FI), namely frequent closures (FCIs) or generators (FGIs), over a stream is already a challenging task, and mining both is even more so. Yet, FGIs and FCIs jointly serve in many practical settings and underlie popular association rule bases. To date, the only way to approach the task is by the (impractical) combination of two dedicated miners. As a remedy, we propose a novel joint miner implementing a holistic approach –emphasizing the underlying equivalence classes– to the joint maintenance of FGIs and FCIs upon stream updates. With a sliding window mining schema, we build upon known results on window expansion and provide novel results on window shrinking while adapting both to the support-based filtering. Thus, through the collaborative maintenance of both FCI and FGI, our method achieves effort factoring that helps it outperform its sole FGI stream mining competitor and keep up with two FCI ones.

A domain ontology (DO) is a machine-readable knowledge repository which, whenever properly exploited, can help to discover meaningful and intelligible patterns from compatible datasets. Yet since such data is naturally graph-shaped, the corresponding task amounts to mining what we call ontologically-generalized graph patterns. We study the underlying problem within a dairy production context where a dedicated DO has been designed beforehand. Two alternative mining approaches have been designed, both representing adaptations of methods from the literature. We evaluated them on an excerpt from our dairy production dataset and report here their respective limitations. We also sketch a way to approach the design of ontology-powered graph miner.

Mining association rules from data streams is a challenging task due to the (typically) limited resources available vs. the large size of the result. Frequent closed itemsets (FCI) enable an efficient first step, yet current FCI stream miners are not optimal on resource consumption, e.g. they store a large number of extra itemsets at an additional cost. In a search for a better storage-efficiency trade-off, we designed Ciclad, an intersection-based sliding-window FCI miner. Leveraging in-depth insights into FCI evolution, it combines minimal storage with quick access. Experimental results indicate Ciclad's memory imprint is much lower and its performances globally better than competitor methods.