AcknowledgementsThis work is partially supported by Project supported by National

10 shows the runtime of all algorithms on dataset kosarok. For dataset kosarok, PrePost+, PrePost, and FIN have almost the same performance, though PrePost+ and PrePost run a little bit faster than FIN. FP-growth∗ perform worst and is lower than other three algorithms by a factor of 2 on average.
Based on the above analysis, we find that PrePost+ always performs best no matter which dataset is used and what the minimum threshold is.