摘要
为及时准确的了解稻米储藏全程质量变化情况,预测储藏品质。对实验室室温储藏大米每15 d检测其脂肪酸值,还原糖、黏度值作训练样本,采用基于反向传播算法的人工神经网络建立了稻米储藏品质的模型,预测大米的脂肪酸值,还原糖和黏度随时间变化的规律,通过试验验证,确定了预测模型具有较高的评价精度、较低的误差率。应用Matlab软件试验,评价准确度达99.8%。
A Back Propagation model of rice storage quality of neural network has been established to precisely realize the rice storage quality alteration conditions promptly.The training sample was fatty acid value,reducing sugar and viscosity of rice,being detected in interval 15 days.The model forecasted fatty acid value,reducing sugar and viscosity of rice of time-dependent rule.Predictive model which had higher estimate precision and lower error rate has been determined by experimental testing.The experimentation which uses Matlab software has estimated a accuracy rating reaching 99.8％.
出处
《中国粮油学报》
EI
CAS
CSCD
北大核心
2014年第4期108-112,共5页
Journal of the Chinese Cereals and Oils Association
基金
黑龙江省自然科学基金重点项目(ZJN0505-01)
黑龙江省科技公关重点项目(GB05C10101)
关键词
BP神经网络
预测分析
脂肪酸值
还原糖
黏度
BP Neural network
estimate analysis
fatty acid value
reducing sugar
viscosity
