(中南大學(xué) 材料科學(xué)與工程學(xué)院，長(zhǎng)沙 410083)
摘 要: 用偏最小二乘法(PLS)結(jié)合反向傳播人工神經(jīng)網(wǎng)絡(luò)(BPN)方法對(duì)7005鋁合金力學(xué)性能與工藝參數(shù)之間的關(guān)系進(jìn)行定性分析和計(jì)算。結(jié)果表明：用PLS法對(duì)實(shí)驗(yàn)數(shù)據(jù)作模式識(shí)別優(yōu)化處理的結(jié)果與實(shí)驗(yàn)很吻合，能夠指明該合金工藝參數(shù)優(yōu)化的方向；用BPN定量計(jì)算的結(jié)果與實(shí)驗(yàn)測(cè)定值符合也較好；將PLS與BPN法有機(jī)地聯(lián)系起來(lái)，有利于克服過(guò)擬合，提高BPN預(yù)報(bào)的準(zhǔn)確性。用留一(LOO)交叉驗(yàn)證法分別對(duì)3種模型PLS、BPN和PLS-BPN的合金性能預(yù)報(bào)結(jié)果進(jìn)行驗(yàn)證，其中PLS-BPN模型預(yù)測(cè)的均方根誤差(RMSE)和平均相對(duì)誤差(MRE)均最低，更適合于7005鋁合金性能預(yù)報(bào)。
關(guān)鍵字: 7005鋁合金；偏最小二乘法(PLS)；神經(jīng)網(wǎng)絡(luò)(BPN)；PLS-BPN；留一(LOO)交叉
（School of Materials Science and Engineering, Central South University,Changsha 410083, China）
Abstract:The mechanical properties of 7005 Al alloys were qualitatively analyzed by partial least squares(PLS) method and quantitatively calculated by using back propagation artificial neural network(BPN) with the same processing parameters as features. The calculated results are in agreement with experimental ones basically. In order to solve the overfitting problem, a novel method hybridizing PLS and BPN to forecast the mechanical properties of the alloys was proposed and tested. PLS method can compute the scores for the principal components according to the sorts of components and thus compress the input data for BPN with linear arithmetic. The forecasting performances were compared with each other by using leave-one-cross-validation(LOOCV) among three models, i.e. the hybrid model(PLS-BPN), BPN model and PLS model. The root mean squared error (RMSE) and the mean absolute relative error (MRE) of PLS-BPN are the lowest. Consequently, the hybrid model is more suitable to forecast the mechanical properties of the alloys.
Key words: 7005Al alloys; partial least squares(PLS); back propagation network(BPN); PLS-BPN; leave-one-cross (LOOC)
