针对长江中下游稻区不同粳稻品种,以不同纬度多个地点的860份稻米样本为材料,基于11项食味品质指标,运用主成分分析、多元线性回归、偏最小二乘法、判别分析、分类回归树(classification and regression free,CART)决策树与反向传播(bac...针对长江中下游稻区不同粳稻品种,以不同纬度多个地点的860份稻米样本为材料,基于11项食味品质指标,运用主成分分析、多元线性回归、偏最小二乘法、判别分析、分类回归树(classification and regression free,CART)决策树与反向传播(back propagation,BP)神经网络分别建立稻米食味品质等级预测模型,通过验证模型的准确性与稳定性,筛选出适合稻米食味品质综合评价的方法或模型。结果表明:BP神经网络模型的2次验证准确度与稳定性均最高,预测准确率分别为92.68%与92.31%;偏最小二乘法模型的2次预测准确率均在80%以上(分别为80.49%、87.18%),但2次验证结果相差较大(6.69%);判别分析与多元线性回归模型的平均预测准确率相近(分别为80.12%、78.77%),但判别分析模型的稳定性优于多元线性回归模型;主成分分析模型的平均预测准确率最低(67.32%),且2次验证结果差异也较大(8.94%);CART决策树模型的稳定性最差,2次验证准确率分别为53.66%与89.74%,相差达36.08%。因此,利用BP神经网络模型预测稻米食味品质等级具有较高的准确性与稳定性,可为长江中下游稻区稻米食味品质综合评价与优质食味水稻品种的筛选提供理论与方法支持。展开更多
Ti-6A1-4V has a wide range of applications, especially in the aerospace field;however, it is a difficultto- cut material. In order to achieve sustainable machining of Ti?6A1-4V, multiple objectives considering not onl...Ti-6A1-4V has a wide range of applications, especially in the aerospace field;however, it is a difficultto- cut material. In order to achieve sustainable machining of Ti?6A1-4V, multiple objectives considering not only economic and technical requirements but also the environmental requirement need to be optimized simultaneously. In this work, the optimization design of process parameters such as type of inserts, feed rate, and depth of cut for Ti-6A1-4V turning under dry condition was investigated experimentally. The major performance indexes chosen to evaluate this sustainable process were radial thrust, cutting power, and coefficient of friction at the toolchip interface. Considering the nonlinearity between the various objectives, grey relational analysis (GRA) was first performed to transform these indexes into the corresponding grey relational coefficients, and then kernel principal component analysis (KPCA) was applied to extract the kernel principal components and determine the corresponding weights which showed their relative importance. Eventually, kernel grey relational grade (KGRG) was proposed as the optimization criterion to identify the optimal combination of process parameters. The results of the range analysis show that the depth of cut has the most significant effect, followed by the feed rate and type of inserts. Confirmation tests clearly show that the modified method combining GRA with KPCA outperforms the traditional GRA method with equal weights and the hybrid method based on GRA and PCA.展开更多
文摘针对长江中下游稻区不同粳稻品种,以不同纬度多个地点的860份稻米样本为材料,基于11项食味品质指标,运用主成分分析、多元线性回归、偏最小二乘法、判别分析、分类回归树(classification and regression free,CART)决策树与反向传播(back propagation,BP)神经网络分别建立稻米食味品质等级预测模型,通过验证模型的准确性与稳定性,筛选出适合稻米食味品质综合评价的方法或模型。结果表明:BP神经网络模型的2次验证准确度与稳定性均最高,预测准确率分别为92.68%与92.31%;偏最小二乘法模型的2次预测准确率均在80%以上(分别为80.49%、87.18%),但2次验证结果相差较大(6.69%);判别分析与多元线性回归模型的平均预测准确率相近(分别为80.12%、78.77%),但判别分析模型的稳定性优于多元线性回归模型;主成分分析模型的平均预测准确率最低(67.32%),且2次验证结果差异也较大(8.94%);CART决策树模型的稳定性最差,2次验证准确率分别为53.66%与89.74%,相差达36.08%。因此,利用BP神经网络模型预测稻米食味品质等级具有较高的准确性与稳定性,可为长江中下游稻区稻米食味品质综合评价与优质食味水稻品种的筛选提供理论与方法支持。
文摘Ti-6A1-4V has a wide range of applications, especially in the aerospace field;however, it is a difficultto- cut material. In order to achieve sustainable machining of Ti?6A1-4V, multiple objectives considering not only economic and technical requirements but also the environmental requirement need to be optimized simultaneously. In this work, the optimization design of process parameters such as type of inserts, feed rate, and depth of cut for Ti-6A1-4V turning under dry condition was investigated experimentally. The major performance indexes chosen to evaluate this sustainable process were radial thrust, cutting power, and coefficient of friction at the toolchip interface. Considering the nonlinearity between the various objectives, grey relational analysis (GRA) was first performed to transform these indexes into the corresponding grey relational coefficients, and then kernel principal component analysis (KPCA) was applied to extract the kernel principal components and determine the corresponding weights which showed their relative importance. Eventually, kernel grey relational grade (KGRG) was proposed as the optimization criterion to identify the optimal combination of process parameters. The results of the range analysis show that the depth of cut has the most significant effect, followed by the feed rate and type of inserts. Confirmation tests clearly show that the modified method combining GRA with KPCA outperforms the traditional GRA method with equal weights and the hybrid method based on GRA and PCA.
