一種基于誤差變化率的自適應(yīng)反向傳播算法
AN ADAPTIVE BACKPROPAGATION ALGORITHM BASED ON ERROR RATE OF CHANGE
-
摘要: 本文針對BP算法存在收斂速度慢的缺點,提出一種基于網(wǎng)絡(luò)動態(tài)訓練誤差變化率自動校正學習步長和沖量因子的自適應(yīng)反向傳播算法。異或問題的仿真結(jié)果表明,該方法具有較快的收斂速度。Abstract: In order to overcome slow convergence rate of the standard BP algorithm, this paper presents an adaptive backpropagation algorithm which can update learning rate and birr factor automatically based on dynamic training error rate of change. Simulation result of the XOR problem shows much faster convergence rate can be obtained by this algorithm.
-
管霖,程時杰,陳德樹.適用于控制型神經(jīng)網(wǎng)絡(luò)的快速學習算法.華中理工大學學報,1995, 23(4): 29-33.[2]張建偉.模糊神經(jīng)網(wǎng)絡(luò)TS模糊模型辨識及其在預(yù)測控制中的應(yīng)用:[碩士學位論文].太原:太原工業(yè)大學自動化系,1997,4.[3]孫德保,高超.一種實用的克服局部極小的BP算法.信息與控制,1995, 24(5): 284-287. -
計量
- 文章訪問數(shù): 2304
- HTML全文瀏覽量: 141
- PDF下載量: 421
- 被引次數(shù): 0
下載:
