基于PCA-ARIMA模型的高血压发病率预测

目的高血压发病率是政府和相关医学工作者预防和监测高血压的重要依据之一。方法利用主成分分析(principal component analysis,PCA)对因子进行线性筛选,获得保留因子后利用ARIMA进行建模预测,即为PCA-ARIMA多维时间序列组合预测模型。结果高血压发病率的拟合与独立预测结果表明,PCA-ARIMA优于PCA-MLR、ARIMA等参比模型。结论本文提出的基于主成分分析和ARIMA模型(PCA-ARIMA模型)的建模有助于提高模型的预测精度。     

Seasonal patterns of dispersal of ascospores of Cryphonectria parasitica (chestnut blight)

Infected barks of chestnut blight cankers, caused by Cryphonectria parasitica , were collected from a naturally infected orchard and incubated at different temperatures. Cankers started to discharge ascospores about a week after incubation at 15–25°C; most ascospores were collected at 20 and 25°C. When incubated at 5, 10 or 30°C, only a few cankers released a small number of ascospores and only during the later stages of incubation. However, the rate of formation of perithecia was not affected by the incubation temperature. The number of airborne ascospores was monitored using a volumetric spore trap in a chestnut orchard during 1996 and 1997. In both years, the number of ascospores trapped daily varied greatly, but in general it increased sharply from March onwards, reached a peak in May, and then declined steeply. There was a significant correlation between daily counts of ascospores and air temperature. Time-series transfer function (TF) analysis showed a positive association of the daily number of ascospores with increasing temperature, rain events and wet/humid conditions. In general, values predicted by the TF model agreed well with the observed pattern. However, a multiple regression equation based on TF analysis failed to provide a satisfactory prediction of the daily number of ascospores.     

变桨距风力发电系统模糊控制

提出了一种基于灰色-ARIMA的金融时间序列智能混合预测模型。首先建立金融时间序列灰色预测模型,并采用PSO算法对灰色模型的三个参数进行优化;利用ARIMA算法对预测模型的残差进行分析,同时采用遗传算法对ARIMA的系数进行优化;最后用ARIMA的残差预测结果对灰色预测模型进行补偿。结果表明,以较好的精度拟合一段时期内MA＜107的时间序列,预测误差控制在5%以上,与单纯的灰色预测算法和神经网络算法相比,在平均绝对误差、均方根误差和趋势准确率三项评价指标上,具有明显优势。     

科技成果转化过程分析

选取2002~2013年我国石油进出口贸易量的数据进行建模分析。首先运用小波分析理论将贸易量数据进行分解,识别出数据的主要特征和细节特征,针对不同特征进行识别和平稳性检测和参数估计,建立相应的ARIMA模型,并进行预测加权合成。仿真结果表明,小波分析结合ARIMA组合模型的预测精度远远大于为改进的ARIMA预测模型,从而为科学合理的决策提供更为精确的预测模型。     

基于条件植被温度指数的冬小麦产量预测

条件植被温度指数(VTCI)综合了地表主要参数——植被指数(NDVI)和地表温度(LST),能够较为准确地对干旱进行监测,可为抗旱救灾、遥感作物估产等提供科学依据。在改进层次分析法的加权VTCI与冬小麦产量的相关性研究成果和VTCI的季节性ARIMA模型干旱预测研究成果基础上,对关中平原的冬小麦产量进行向前1旬、2旬和3旬的预测研究。研究结果表明,产量预测结果与产量监测结果吻合较好,预测精度随着预测步长的增大而降低,关中平原4个地级市平均产量预测结果的最大相对误差为3.27%,说明用该方法可以进行向前3旬的产量预测。     

海拉尔地区降水量的ARIMA乘积季节模型研究

1个地区降水量的变化受到诸多因素的影响,其既有长期趋势、季节效应,又具有随机扰动效应。在降水量的分析中,找出所有影响降水量的变化因素是十分困难的,因此可以利用降水量序列的自身变化来寻找序列中蕴含的发展规律。本文以海拉尔地区1957年1月~2007年8月逐月平均降水量数据为基础,应用SAS统计软件经过多次拟合优选,最终建立了适合海拉尔地区月平均降水量的AR IMA(p,d,q)×(P,D,Q)s乘积模型,并利用该模型对海拉尔地区2007年的降水量进行了预测。     

我国大豆加工企业的动态套期保值业务模式

为提高我国大豆加工企业的价格风险管理水平提供理论参考,基于套期保值理论,采用ARIMA模型和VAR模型对国内大豆和豆油的价格与基差变动进行预测,提出大豆加工企业的动态套期保值业务模式。结论:大豆加工企业应首先根据未来大豆和豆油价格变化趋势的预测,确定是否存在价格波动风险并决定是否进行套期保值;其次,根据现货和期货市场的价格波动情况确定合适的套期保值比率;最后,根据预测的基差变化趋势,动态调整套期保值的交易策略。     

长江靖江段沿岸似鳊的时间格局及生长特征

为了解长江下游沿岸似鳊（Pseudobrama simoni）的种群变动及生长状况,研究了2002-2009年间在靖江沿岸用定置张网采集的每月2次或3次共237份渔获物样本。结果表明：237份样本共有鱼类73 960尾、409 256.23 g。其中似鳊4 253尾、39 618.05 g,分别占总渔获量的5.75%和9.68%。平均每样本19尾、180.08 g,最高达145尾、1714.6 g。时间格局分析表明,似鳊月渔获重量呈非平稳的随机过程,对2002-2009年的月渔获重量进行ARIMA建模拟合,建立了方程为 (1－0.387B)（1－B¹²）lny_t=（1－0.555B¹²）e_t的ARIMA（1,0,0）（0,1,1）₁₂的预测模型,该模型对2003-2009年的预测精度达83.21 %～93.90%。对2005年采集的全部个体作性别、年龄和生长特征分析,显示似鳊的雌雄性比为1∶2.3。雄性有1～3龄3个年龄组,雌性只有1～2龄2个年龄组,年龄结构明显低于姚江和钱塘江种群。雄性的Von bertalanffy生长方程为：体长L_t=147.17×［1－e^{－0.1648×(t+3.2036)}］,体重W_t=44.36×［1－e^{－0.1648×(t+3.2036)}］^2.785;推算的渐近体长L_∞为147.17 mm,渐近体重W_∞为44.36 g。     

Optimal forest rotation periods: integrating timber production and carbon sequestration benefits in Pinus tabulaeformis plantations on the Loess Plateau,P.R. China

Determining the optimal rotation period was a crucial component of forest sustainable management strategies, especially under climate change. This paper had two objectives: (1) to determine the economic benefits and optimal rotation periods for timber production when coupled to carbon sequestration, as predicted by time series prediction models for Pinus tabulaeformis plantations in China; and (2) to evaluate how different carbon prices and interest rates affected optimal rotation periods using the forest land expectation value. The results suggested that time series prediction models were valuable for estimating timber volumes and carbon sequestrations based on surveys of different-aged stands. Importantly, since integrating carbon sequestrations into timber production benefits did not increase optimal rotation periods, this should promote P. tabulaeformis plantation management. In the sensitivity analysis, a higher carbon price increased the profitability of carbon sequestration and timber production, but not optimal rotation periods, though they were reduced under higher interest rates. In conclusion, incorporating both timber production and carbon sequestration benefits would sharply increase forest-based revenues, while realizing the carbon sequestration potential of P. tabulaeformis plantations. This approach was clearly useful to the development of reforestation/afforestation projects trying to mitigate climate change and also provided a theoretical basis for sustainable forest management.     

银杏落叶期叶片NDVI值2种拟合方法比较

【目的】对银杏Ginkgo biloba L.落叶期(9—12月)的叶片进行连续定期定点的高光谱测量,计算出能代表其生长状况和营养信息的高光谱参量NDVI值,通过选用两种不同的曲线拟合方法对其NDVI值进行宏观上的以时间为自变量的曲线拟合,选出最优拟合方法,更好地了解银杏落叶期叶片光谱特征参量NDVI值的变化趋势,从而更有效地对其进行决策和控制,为植被的大尺度遥感动态监测提供方法参考。【方法】利用SVC HR-1024I全波段地物光谱仪,选取三株健康、生长环境相同、长势相近的中龄银杏为叶片采集对象,对其落叶期冠层叶片进行定期定点定方位的高光谱观测。对获取的高光谱原始数据进行数据筛选与预处理后,通过计算得出叶片的NDVI值,分别采用二次函数拟合法和ARIMA时间序列拟合法对落叶期叶片的NDVI值进行曲线拟合,并对两种拟合方法的拟合结果进行比较,选出最适合银杏叶片落叶期NDVI值的拟合方法。【结果】二次函数拟合结果为NDVI=-0.0221T2+0.0547T+0.711,决定系数R2为0.926,但因拟合结果t值不显著,样本结果随机性大,不具广泛性;ARIMA时间序列拟合中ARIMA(2,1,2)模型估测结果与实际情况接近,R2为0.811,拟合效果较好。【结论】ARIMA时间序列拟合方法比二次行数拟合方法更适用于对银杏落叶期叶片的NDVI值进行拟合。