Research on a Learning Algorithm as Inverse Deduction
Jianping Chen · Journal of Nantong University · 2009
This paper proposes a new algorithm called LAID(learning algorithm as inverse deduction) considering the difficulty of traditional induction learning.Inverse deduction is an approach in which induction learning is treated as the inverse of deduction.The key issues of LAID include the inverse deduction rules,lattice structures and state operators from the perspectives of learning tasks,mode declarations and heuristic information.A learning and reasoning system is implemented using the LAID.Experimental results show that LAID is a feasible and effective algorithm that can generate rules with strong generalization ability.