This chapter compares the results of the data analysis above , and finds that the passed explanatory variables in the research sampling all the listed companies tend to exceed those in the research sampling the listed companies of different professions in numbers , the r2 and adj . r2 in the research sampling all the listed companies tend to be less than those in the research sampling the listed companies of different professions 本章比较分析上述数据处理结果，发现以全体上市公司为样本研究时通过的解释变量数一般多于以各个行业的上市公司为样本研究时通过的解释变量数，以全体上市公司为样本研究时得到的r ~ 2和r ~ 2修正值一般小于以各个行业的上市公司为样本研究时得到的r ~ 2和r ~ 2修正值。
In this paper , the upper - class intelligent control scheme is adopted according to the characteristics of tobacco re - dryer and the present control condition in order to solve the problem that the output tobacco moisture can ? keep stable . that is , to keep the old every pid control link , to add the upper - class computer so that sets up the upper - class intelligent control system that consists of the description management sub - system , the fuzzy compensation on line sub - system and the self - learning system . the following are control procedure : first of all , a series of output values of description can be obtained by the forward inference of basic controller on the basis of the information of input tobacco , then , these output values can be adjusted by the fuzzy compensation controller with inquiring the fuzzy control - list on line in the light of the output tobacco moisture error 本文针对烤机的工艺特点和目前控制现状，提出用上位智能控制的方法解决烤机目前存在的问题，即在保留原有各段pid控制的基础上，增加上位计算机，建立起由配方管理子系统、模糊在线修正子系统和自学习系统组成的上位计算机监控、决策的智能控制系统；控制过程为：基本配方控制器根据来料信息正向推理提供烟叶复烤机各段参数设定值的基值；模糊补偿控制器根据烟叶水分偏差在线查询模糊控制表提供相应参数设定值的修正值，实现烟叶复烤机参数的在线修正；学习环节的引入实现烟叶配方的在线优化。
