Mining Frequent Function Set Based on Gene Expression Programming
Jia Xiao · Chinese Journal of Computers · 2005
Function Mining aims at discovering valid functions from observation data. However, the traditional Function Mining aims at single function, and hence it is difficult to process in complex data set. To solve this problem, this paper proposes a new concept called Frequent Function Set (FFS) with powerful describing ability, presents and implements a new approach named Frequent Function Set Mining (FFSM) to mine FFS based on Gene Expression Programming, and improves the success probability of in FFSM by a new strategy named Precision Threshold Queue (PTQ). Extensive experiments demonstrate the power of FFS and of PTQ that it improves the success-probability by 55 times for mining complex function with high precision.