摘要
从车型、发动机、变速器和轮胎等方面分析了与汽车燃油消耗相关的影响因素,通过主成分方法对影响汽车燃油消耗的变量进行了压缩,消除了各变量之间的线性相关性。再利用BP神经网络对主成分的得分进行预测,建立燃油消耗预测模型。结果表明,与传统BP神经网络相比,采用主成分分析与神经网络相结合的燃油消耗预测模型简化了神经网络结构,提高了预测精度,为预测汽车燃油消耗量提供了新的思路。
Currently, forecast of vehicle fuel consumption has more practical economic significance and value for automobile manufacturers and consumers. In the paper, the influence factors on vehicle fuel consumption are analyzed such as vehicle model, engine, transmission and tire etc, the variables affecting fuel consumption are compressed by principal components for eliminating linear relevance correlation of variables. Principal component scores are predicted by BP neural network and the fuel consumption forecast model is established. The results show that compared with traditional BP neural network,using principal component analysis and neural network combining prediction model of fuel consumption can simplify the model structure, improve prediction precision and provide new ideas for the fuel consumption forecast of the vehicles.
出处
《农业装备与车辆工程》
2015年第6期47-52,共6页
Agricultural Equipment & Vehicle Engineering
关键词
影响因素
主成分分析
BP神经网络
燃油消耗预测
influence factors
principal component analysis
BP neural network
forecast of fuel consumption
