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
Three cultivars of Japonica rice:hybrid Japonica rice,inbred Japonica rice and soft Japonica rice,were used to study the effects of recommended nitrogen fertilizer on grain quality including milling,appearance,nutriti...Three cultivars of Japonica rice:hybrid Japonica rice,inbred Japonica rice and soft Japonica rice,were used to study the effects of recommended nitrogen fertilizer on grain quality including milling,appearance,nutrition,cooking and eating quality,pasting property and their relationships in direct seeding rice.Recommended nitrogen fertilizer improved the brown rice rate,the milled rice rate,and the protein content;increased the chalky grain rate and the chalkiness,and reduced the amylose content,palatability and pasting viscosity.The nitrogen fertilizer caused the appearance,the cooking and the eating quality to deteriorate.Lower amylose content and protein content were found in soft Japonica rice compared to those in hybrid Japonica or inbred Japonica rice under the condition of the recommended rate of nitrogen fertilizer,contributing to high cooking and eating quality.Correlation analysis indicated that palatability was negatively correlated with amylose content,protein content,hardness and gelatinization temperature,and was positively correlated with stickiness,peak viscosity and breakdown.Therefore,the decreased amylose and protein content of the soft Japonica rice cultivar planted under the condition of the recommended rate of nitrogen fertilizer is beneficial to achieving high taste quality.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Three cultivars of Japonica rice:hybrid Japonica rice,inbred Japonica rice and soft Japonica rice,were used to study the effects of recommended nitrogen fertilizer on grain quality including milling,appearance,nutrition,cooking and eating quality,pasting property and their relationships in direct seeding rice.Recommended nitrogen fertilizer improved the brown rice rate,the milled rice rate,and the protein content;increased the chalky grain rate and the chalkiness,and reduced the amylose content,palatability and pasting viscosity.The nitrogen fertilizer caused the appearance,the cooking and the eating quality to deteriorate.Lower amylose content and protein content were found in soft Japonica rice compared to those in hybrid Japonica or inbred Japonica rice under the condition of the recommended rate of nitrogen fertilizer,contributing to high cooking and eating quality.Correlation analysis indicated that palatability was negatively correlated with amylose content,protein content,hardness and gelatinization temperature,and was positively correlated with stickiness,peak viscosity and breakdown.Therefore,the decreased amylose and protein content of the soft Japonica rice cultivar planted under the condition of the recommended rate of nitrogen fertilizer is beneficial to achieving high taste quality.
