基于共轭先验分布的贝叶斯网络分类模型-控制与决策.PDFVIP

绗?7 鍗 绗? 鏈 鎺 鍒 涓 鍐 绛 2012 骞 9 鏈 Vol. 27 No. 9 Control and Decision Sep. 2012 鏂囩珷缂栧彿: 1001-0920 (2012) 09-1393-04 鍩轰簬鍏辫江鍏堥獙鍒嗗竷鐨勮礉鍙舵柉缃戠粶鍒嗙被妯″瀷 鏉ㄩ娑? 鐜嬭穬閽? 閭撳崼寮? 寰愭椽娑 (绗簩鐐叺宸ョ▼瀛﹂櫌 304 鏁欑爺瀹わ紝瑗垮畨710025) 鎽 瑕? 閽堝璐濆彾鏂綉缁滃悗楠屾鐜囬渶璁＄畻鏍锋湰杈归檯鍒嗗竷, 璁＄畻浠ｄ环澶х殑闂, 灏嗗叡杞厛楠屽垎甯冩€濇兂寮曞叆璐濆彾鏂垎绫? 鎻愬嚭浜嗗熀浜庡叡杞厛楠屽垎甯冪殑璐濆彾鏂綉缁滃垎绫绘ā鍨? 閽堝闈炲尯闂寸鏁ｆ牱鏈? 鎻愬嚭涓€绉嶈嚜閫傚簲鐨勬牱鏈鏁ｆ柟娉? 灏嗗皬娉㈠寘 鎻愬彇妯℃嫙鐢佃矾鏁呴殰鐗瑰緛绂绘暎鍖栦綔涓哄垎绫绘ā鍨嬪睘鎬? 浠跨湡楠岃瘉琛ㄦ槑, 妯″瀷鍒嗙被鏁堟灉杈冨ソ, 绠楁硶杩愯閫熷害寰椾互鎻愰珮, 涔熷彲 搴旂敤浜庤繛缁牱鏈拰澶氬垎绫荤殑鎯呭喌, 鎵╁睍浜嗚礉鍙舵柉缃戠粶鍒嗙被鐨勫簲鐢ㄨ寖鍥? 鍏抽敭璇岤 璐濆彾鏂綉锛涘叡杞厛楠屽垎甯冿紱杈归檯鍒嗗竷锛涙ā鎷熺數璺紱鏁呴殰璇婃柇 涓浘鍒嗙被鍙耳 TP181 鏂囩尞鏍囧織鐮? A Bayesian network classi铿乪r based on conjugate prior distribution YANG Ying-tao, WANG Yue-gang, DENG Wei-qiang, XU Hong-tao (The 304 Staff Room锛孴he Second Artillery Engineering College 锛孹i鈥檃n 710025 锛孋hina 锛嶤orrespondent 锛歒ANG Ying- tao 锛孍-mail 锛歽yingtao@163.com) Abstract 锛氾細锛欼n order to reducing calculate costs of Bayesian network, when calculating posterior probability of samples that need the marginal distribution, an approach of Bayesian network classi铿乪r based on conjugate prior distribution is proposed. An adaptive discretization method is also proposed to discrete non-interval samples. The fault feature of analog circuit extracted by wavelet packet is taken as a discrete property of Bayesian network classi铿乧ation model. The simulation result shows that, this classi铿乪r has high accuracy and ef铿乧iency of analog circuit fault diagnosis, and can be applied to continuous and multi-classi铿乧ation case, which extends the scope of application of Bayesian network classi铿乧ation. Key words 锛氾細锛欱ayesian network 锛沜onjugate prior distribution 锛沵arginal distribution 锛沘nalog circuit 锛沠ault diag