基于Lasso回归和BP神经网络的蔬菜短期价格预测组合模型研究
喻沩舸, 吴华瑞, 彭程

Short-Term Price Forecast of Vegetables Based on Combination Model of Lasso Regression Method and BP Neural Network
Weige YU, Huarui WU, Cheng PENG
表1 Lasso回归模型变量选择与参数估计
Table 1 Variable selection and parameter estimation of Lasso regression model
系数影响因素12345678
β1播种面积0-0.28212-0.20547-0.18647-0.19540-0.20108-0.20273-0.20646
β2亩产量000-0.18932-0.29021-0.1019800
β3受灾面积000.4087340.4389490.4454820.5648590.5921840.531208
β4土地成本00.2394790.2908500.3038290.3139820.3602150.3474980.324921
β5物质与服务费用0000.1015410.1075320.1401980.0921920
β6用工成本0000.3863230.389201000
β7城镇居民人均可支配收入000.5232940.5187130.5147190.5902340.3640210.393297
β8城镇居民人口数量00.3105830.2108400.2009870.112749000
β9城镇蔬菜消费价格指数00.4123850.3548170.3123900.3789740.3762130.3740850.375532
β10叶类蔬菜均价000.5701970.5427600.4235700.43750200
β11根茎类蔬菜均价00.4537500.4210970.3923750.3219370.29137400
β12高速公路过路费0000.2108700.1247910.0213970.0787300
β13燃油附加费0000.3430190.3413090.3510350.2531080.271075
β14司机劳务费0000.0392180.0329570.0239850.0932130
β15摊位费0000.1966570.2078660.2512240.2433170.213077
β16人工费0000.2175530.233591000
β17损耗费00.3534110.3289290.2936650.2277550.23680100
β18包装加工费000.2179170.3975910.3193270.3431090.3275190.247327
β19国家经济发展水平000.1530170.0430920.0397590.0431970.0221740
β20人民币汇率000.0128740.0238570000
β21通货膨胀率000.5653250.6101280.5401980.5312080.5310930.501098
β22黄瓜产业扶持金额0000-0.320180-0.290180-0.341820-0.351080
β23气温平均值0-0.43296-0.532850-0.624810-0.619750-0.603290-0.593820-0.592190
β24气温偏离正常值00.4310280.4130980.3230930.4139020.5321010.6321810.491273
AIC3.2287961.9678900.8913440.2457760.1023550.0831230.071226