本文針對(duì)工業(yè)系統(tǒng)運(yùn)行狀態(tài)設(shè)計(jì)評(píng)估方法。首先以整個(gè)系統(tǒng)中某個(gè)子系統(tǒng)的傳感器組
群信號(hào)作為輸入，訓(xùn)練自適應(yīng)增強(qiáng)(Adaboost)分類器，得到該子系統(tǒng)多種假定工作狀態(tài)分類結(jié)果。之后以子系統(tǒng)狀態(tài)分類結(jié)果作為貝葉斯決策(Bayesian Decision)的依據(jù)，運(yùn)用最小錯(cuò)誤率(Minimum Error Ratio)原理估計(jì)整個(gè)系統(tǒng)狀態(tài)。實(shí)驗(yàn)采用仿真形式，以多組模擬傳感器信號(hào)作為輸入，完成系統(tǒng)評(píng)估過(guò)程。
關(guān)鍵詞：自適應(yīng)增強(qiáng)貝葉斯決策系統(tǒng)狀態(tài)評(píng)估最小錯(cuò)誤率

System State Assessment Basted on Adaboost and Bayesian Decision HOU YI-MIN, LUN XIANG-MIN School of Automation Engineering, Northeast Dianli University, 132012, China
Abstract The paper designed a industrial system state assessment method. The sensors’ signal of a sub-system of the whole system is employed as the inputs which are used to train the Adaboost classifier. And the sub-system state is obtained using the classifier. The states of all the sub-system are used as the inputs of the Bayesian Decision. In the experiments, simulation sensors’ signal is used as the inputs to complete the assessment procedure.
Key Words: Adaboost, Bayesian decision, System state assessment, Minimum error ratio