人工免疫系统在手写体数字识别中的应用

基于生物免疫系统内在的模式识别与记忆能力,通过对手写体数字识别问题的研究,首次将借鉴生物免疫系统发展起来的人工免疫机制和算法引入到该问题中来,定义了相关免疫机制的数学表达,设计了具有学习、记忆、自适应和多样性等特性的人工免疫系统模型,并给出了相应的算法.通过对自采集的手写体数字样本库进行的仿真试验,结果表明用于手写体数字识别的人工免疫系统模型是可靠的,其对应算法可操作性强,只要相关参数选择得当,系统能够得到较好的识别效果,从而体现了人工免疫系统在模式识别领域的优越性和技术潜力.最后对下一步的研究工作进行了展望.

Application of Artificial Immune System in Hand-written Number Recognition

Based on the inherent pattern recognition and memory ability of biology immune system, with the research of hand - written number recognition, the artificial immune mechanism and algorithm advanced with biology immune system are introduced into this problem for the first time. The mathernatical expressions of the corresponding immune mechanism are defined and the artificial immune system model with abilities of study, memory and self - a- dapting and diversity is designed, at the same time, the corresponding algorithm of the model is given. The simulation experiment on the self collective hand -written number sample set is done, and the result shows that the artificial immune system model used for the hand - written number recognition is reliable and the corresponding algorithm is operated easily, and if the relative variables could be chosen properly, better recognition effect of the system would be gotten. So, it indicates that the artificial immune system has advantages and technological potential in the field of pattern recognition. At last, the next research work is previewed.