Towards a Particle Swarm Optimization-Based Regression Rule Miner
Bart Minnaert, David Martens · 2012
We present the work in progress on a rule mining algorithm for regression using particle swarm optimization (PSO). Sub problems occuring during development involve the encoding of rules as particles and suitable PSO parameter tuning. A key subtask is the selection of a good rule learning heuristic. We introduce a novel heuristic for which preliminary results show promise.