基于非参数核密度估计法的红枣市场价格波动与风险评估

Fluctuation and risk evaluation of jujube market price based on the non-parametric kernel density estimation method

利用非参数核密度方法,以5种红枣(灰枣、骏枣、金丝枣、青枣和冬枣)的批发市场价格为研究对象,比较分析不同红枣的短期价格波动形态及波动强度,并对其短期价格波动风险进行评估和比较。结果表明:从波动形态看,总体上红枣市场价格短期波动有明显谷峰和谷底,呈"春冬高、夏秋低"季节性特征,其中骏枣和灰枣的价格呈"V"型波动,金丝枣和冬枣的价格呈"M"和"W"型波动,青枣价格呈"N"型波动;从波动强度看,整体上鲜食枣价格波动强度高于干制枣,大小排序依次为冬枣>青枣>金丝枣>骏枣>灰枣;红枣市场价格波动风险分布不对称,降价风险高于涨价风险,其中冬枣降价风险最高,其次是灰枣和骏枣,金丝枣降价风险略低,只有青枣价格下跌风险与上涨跌风险发生的概率接近一致;从强警风险概率(价格波动幅度在±15%以上)分布看,鲜食枣价格出现强警的概率高于干制枣。

Using non-parametric kernel density method and taking the monthly wholesale prices of 5 kinds of jujubes （Jinsi jujube,Jun jujube,Dong jujube, Hui jujube and Qing jujube） as the object of study, the wave pattern and intensity of short term fluctuation of different kinds of jujubes were compared and analyzed,and the short-term price volatility risk was evaluated and compared. The results showed that： From the point of price fluctuation pattern, in general, the short- term pattern of jujube market price displayed distinct peak and trough,which was high in spring and winter, as well as low in summer and autumn, the wave pattern of Jun-jujube and Hui-jujube market price was ＂V＂ type, and the wave pattern of Dong jujube and Jinsi jujube market price was ＂M＂ and ＂W＂ type, the price of Qing jujube showed ＂N＂ type waver From the view of price fluctuation intensity, in the whole fluctuation intensity of fresh jujube price is much higher than that of dried jujube,the descending order of fluctuating intensity of jujubes is Dong jujube,Qing jujube,Jinsi jujube, Jun juiube, Hui jujube~ the price risk distribution of jujube market is asymmetric and the probability of falling of jujube market price is higher than that of rising, the failing risk of Dong jujube market price is the greatest, followed by Jun jujube and Jinsi jujube,the falling risk of Hui jujube market price is slightly lower,only Qing jujube price rise probability remained stable. Strong alarm probability risk of fresh jujubes price is higher than that of dried jujubes by the distribution of strong alarm probability risk.