在智能通信网的学习过程中,针对传统归纳演绎学习的困难,提出一种逆演绎的学习算法。从学习任务和逻辑程序角度出发,研究逆演绎学习算法的规则、霍恩状态函子和激励信息等问题,并利用逆演绎的学习算法实现了智能通信网的学习推理系统。测试结果表明,该算法能高效地归纳出逻辑程序。
In learning process of intelligence communication network,there is difficulty in traditional induction and deduction learning.To address this,in the paper we propose a new algorithm called LAID(learning algorithm as inverse deduction).The key issues studied in LAID include reverse deduction learning algorithm rule,Horn state operator and incentive information from the perspectives of learning tasks and logic program.The learning and reasoning system of intelligence communication network is implemented using LAID.Experimental results show that LAID can efficiently generalise logic program.