Maximal Mixed-Drove Co-Occurrence Patterns
[ 1 ] Instytut Informatyki, Wydział Informatyki i Telekomunikacji, Politechnika Poznańska | [ P ] pracownik
2021
rozdział w monografii naukowej / referat
angielski
EN Mining of Mixed-Drove Co-occurrence Patterns can be very costly. Widely used, Apriori-based methods consist in finding spatial co-location patterns in each considered timestamp and filtering out patterns that are not time prevalent. Such an approach can be inefficient, especially for datasets that contain co-locations with a high number of elements. To solve this problem we introduce the concept of Maximal Mixed-Drove Co-occurrence Patterns and present new algorithm MAXMDCOP-Miner for finding such patterns. Our experiments performed on synthetic and real world datasets show that MAXDCOP-Miner offers very high performance when discovering patterns both in dense data and for low values of spatial or time prevalence thresholds.
16.08.2021
15 - 29
20
70