| 摘要: |
| ［目的］以顷均农业机械总动力及影响因素为研究对象，量化各因素对顷均总动力的影响强度，进而为定量分析顷均农机总动力影响因素提供途径。［方法］以新疆1990~2014年顷均农业机械总动力及其影响因素作为数据源，建立关系数据模型。构建嵌入式灰色神经网络（Embed grey neural network，EGNN）模型，并与EGNN与平均影响值（Mean impact value，MIV）方法相结合，建立了EGNN-MIV影响因素重要度计算模型，并对新疆25年顷均农机总动力的影响因素MIV进行测度。［结果］新疆1990~2014年顷均农机总动力受各种影响因素的交叉影响。各影响因素标准化后的MIV分别为0.70、0.39、0.39、0.46、0.20、0.50、0.90、0.74。顷均农机总动力影响最大的前3个因素分别为：劳均土地面积、最大种植面积农作物比重以及耕地均顷GDP。对顷均农机总动力的影响最小的因素为农业比较劳动生产率。［结论］建立的EGNN-MIV模型在量化模型输入对输出的影响程度方面有效和可行。顷均农机总动力主要受土地面积分布、农作物种植结构和均顷GDP等因素的显著影响。为定量分析顷均农业机械总动力等时间序列的影响因素提供了思路和方法，同时为农机管理部门提供决策参考。 |
| 关键词: 顷均农机总动力 影响因素 影响强度 神经网络 平均影响值 |
| DOI：10.7621/cjarrp.1005-9121.20171104 |
| 分类号: |
| 基金项目:国家自然科学基金资助项目“基于工况自识别的拖拉机排放特性研究”（51768071） |
|
| STUDY OF IMPACT FACTOR FOR AGRICULTURAL MACHINERY TOTAL POWER BASED ON GREY NEURAL NETWORK AND MIV |
|
Tursun Mamat1, Ding Weimin2,3, Muhammad Hassan1
|
|
1.College of Engineering, Nanjing Agricultural University, Nanjing, Jiangsu 210031, China;2.1. College of Engineering, Nanjing Agricultural University, Nanjing, Jiangsu 210031, China;3.2. School of Mechanical and Traffic Engineering, Xinjiang Agricultural University, Urumqi 830052, China
|
| Abstract: |
| Taking the total power of agricultural machinery for average hectare（TPAMAH） and its influencing factors as the research object, the influence strength of each factor on TPAMAH is quantified, which provided a way to quantitatively analyze the influencing factors of TPAMAH.During the farm mechanization in Xinjiang in the last decade, each impact factor of TPAMAH has different impact in all parts of Xinjiang from 1990~2014. The TPAMAH and its impact factor, the influence intensity of all factors were quantified, and then the way of quantitative analysis of the influencing factors of the total power of agricultural machinery was provided. The embedded grey neural network (EGNN) forecasting model and the EGNN-MIV (mean impact factor) model were used for measuring MIV of each impact factor of TPAMAH. Then MIV was calculated by using EGNN-MIV approach. According to minimum mean square error of the EGNN-MIV model, the optimal MIV was determined. The results showed that TPAMAH was cross influenced by various factors in 1990~2014. Standardization MIV of each impact factor for TPAMAH was 0.70, 0.39, 0.39, 0.46, 0.20, 0.50, 0.90 and 0.74. The first 3 highest impact factors were land area per capita, the proportion of the largest planting area of crops, and GDP of mean area. The agricultural comparative labor productivity had the minimum MIV with standardization value of 0.20. It concluded that the established EGNN-MIV model was effective and feasible in quantifying the impact of model input on output. The results also indicated that the land area per capita, the proportion of the largest planting area of crops and GDP of mean area had significant impact on TPAMAH. This work can provide a reference for the quantification of the impact factors on the TPAMAH and the similar time series. |
| Key words: total power of agricultural machinery for average hectare impact factor influencing intensity neural network mean impact value (MIV) |
