Comparison of Frequent Pattern Mining Algorithms in Internet of Things
Halil Ibrahim Dede, Cemile Timurkaan, Duygu Kurt, Amina Kofrc · 2020
Frequent pattern mining is an important topic that is needed for the Internet of Things (IoT) applications frequently. Many IoT applications have been developed in which contionuous streaming data is used. In this study, performance evaluation of FP-stream (Frequent Pattern-Stream) algorithm and weighted sliding window mining (WMFP-SW) algorithms, which are developed for frequent pattern analysis in IoT, is performed using a novel real world dataset. The dataset was created by collecting 6 different sensor values over a 3 month period in an established testbed. The results showed that FP-stream algorithm is more successful than WMFP-SW algorithm at execution time, whereas WMFP-SW is better at the total number of mining frequent patterns.