The Application of Pearson Correlational Analysis Method in Air Quality Analysis of Beijing-Tianjin-Hebei Region

Based on the analysis of monitoring data on six pollution indexes of SO2, NO2, CO, O3, PM10 and PM2.5 from 53 monitoring points in 7 cities, including Beijing, Tianjin, Shijiazhuang, etc., from April 8 of 2014 to July 23 of 2014, this article adopted Pearson correlation coefficient method to determine the relevance among each pollutant of these cities with the help of SPSS. The results showed that such three leading indexes as SO2, PM10 and PM2.5 had strong correlation in Beijing, Tianjin and main cities of Hebei. Finally, some suggestions and preventive measures for the cooperative governance of air pollution in Beijing-Tianjin-Hebei Region were put forward, hoping this can help them.

Pearson＇s Correlation Coefficient： A More Realistic Threshold for Applications on Autonomous Robotics

Many applications for control of autonomous platform are being developed and one important aspect is the excess of information, frequently redundant, that imposes a great computational cost in data processing. Taking into account the temporal coherence between consecutive frames, the PCC （Pearson＇s Correlation Coefficient） was proposed and applied as： discarding criteria methodology, dynamic power management solution, environment observer method which selects automatically only the regions-of-interest; and taking place in the obstacle avoidance context, as a method for collision risk estimation for vehicles in dynamic and unknown environments. Even if the PCC is a great tool to help the autonomous or semi-autonomous navigation, distortions in the imaging system, pixel noise, slight variations in the object＇s position relative to the camera, and other factors produce a false PCC threshold. Whereas there are homogeneous regions in the image, in order to obtain a more realistic Pearson＇s correlation, we propose to use some prior known environment information.

基于Pearson相关系数与广义S变换的低压直流微电网的故障选线方法

针对低压直流微电网线路的故障选线方法的研究有限,提出一种基于Pearson相关系数与广义S变换的故障选线方法。首先,介绍了低压直流微电网的一般组成。在此基础上,研究了低压直流微电网的单极故障以及极间故障的故障特征,提出选线方法,对各线路正负极首末端电流差进行Pearson相关系数计算来确定故障线路。然后对所选故障线路正负线路首端的电流量进行广义S变换,计算出其能量和的比值判别故障极性或者极间故障。最后,在PSCAD/EMTDC中搭建出低压直流微电网模型以输出各线路电流数据,并利用Matlab对数据进行仿真。结果表明,所提选线方法直接有效,且耐受过渡电阻的能力较强。

Spatio-temporal Variation Characteristics of Extreme Climate Events and Their Teleconnections to Large-scale Ocean-atmospheric Circulation Patterns in Huaihe River Basin,China During 1959–2019

Huaihe River Basin(HRB) is located in China’s north-south climatic transition zone,which is very sensitive to global climate change.Based on the daily maximum temperature,minimum temperature,and precipitation data of 40 meteorological stations and nine monthly large-scale ocean-atmospheric circulation indices data during 1959–2019,we present an assessment of the spatial and temporal variations of extreme temperature and precipitation events in the HRB using nine extreme climate indices,and analyze the teleconnection relationship between extreme climate indices and large-scale ocean-atmospheric circulation indices.The results show that warm extreme indices show a significant(P < 0.05) increasing trend,while cold extreme indices(except for cold spell duration) and diurnal temperature range(DTR) show a significant decreasing trend.Furthermore,all extreme temperature indices show significant mutations during 1959-2019.Spatially,a stronger warming trend occurs in eastern HRB than western HRB,while maximum 5-d precipitation(Rx5day) and rainstorm days(R25) show an increasing trend in the southern,central,and northwestern regions of HRB.Arctic oscillation(AO),Atlantic multidecadal oscillation(AMO),and East Atlantic/Western Russia(EA/WR) have a stronger correlation with extreme climate indices compared to other circulation indices.AO and AMO(EA/WR) exhibit a significant(P < 0.05) negative(positive)correlation with frost days and diurnal temperature range.Extreme warm events are strongly correlated with the variability of AMO and EA/WR in most parts of HRB,while extreme cold events are closely related to the variability of AO and AMO in eastern HRB.In contrast,AMO,AO,and EA/WR show limited impacts on extreme precipitation events in most parts of HRB.

Geostatistical Correlation of Aquifer Potentials in Abia State, South-Eastern Nigeria

In this paper, a collection of statistical correlation methods is used in the study of aquifer potentials in Abia State of south-eastern Nigeria. The Physiology, geomorphology and hydrogeology of the area are first presented. Sixty-six Vertical Electrical Sounding (VES) data sets are used to determine the aquifer. Demographic studies are then carried out in 220 communities in order to determine the relationship between population size on one hand and a unit draw-down of wells due to groundwater abstraction on the other. The relationship between geological Formation, aquifer potentials and depth of boreholes are then calculated using Pearson’s correlation matrix. Results show that the mean population of persons appears to be higher in Bende-Ameki Formation (of Eocene-Oligocene age) and the late Tetiary-Early Quaternary Coastal Plain Sands, than in the Cretaceous shale Formation of Asata Nkporo. The mean population of persons sitting on these Formations is 31,200, 18,370 and 5400 respectively. Furthermore, it is observed that a population increase of about 50 persons in a community in Abia State is accompanied by a unit volume (1 m3) draw-down of wells due to groundwater abstraction. It is therefore concluded that population size is positively correlated with groundwater abstraction, aquifer potentials and geological Formation favouring aquifer in Abia State.

基于LSTM神经网络的乡村非遗发展状况预测

目前,乡村非物质文化遗产(以下简称非遗)开始进入大众视野。为加快其传承与发展的进程,本文对乡村非遗文化的发展做出了预测。首先对收集到的数据进行信效度检验,并通过皮尔逊相关性分析及多元线性回归分析得出了影响非遗发展的三个关键因素,同时得出了影响非遗文化发展前景的回归方程。其次,分别从关注程度、政府方面以及民间组织方面利用LSTM神经网络对乡村非遗文化发展进行预测,得到相应的预测结果,即:未来三年乡村非遗文化的发展总体呈上升趋势。最后,基于预测结果,为各地乡村政府提供了相应的改进建议,以促进非遗文化的交流与传播。

基于Pearson评价法的室内污染源辨识研究

目的从室内污染物的时空分布特性出发,提出基于Pearson评价法的室内污染源辨识策略,从而快速准确辨识室内污染源.方法以Pearson相关系数作为辨识指标以及得分率作为辨识结果的评价标准,先采用控制变量法对Pearson评价法的辨识有效性进行验证,作为后续研究开展的基础,再基于Pearson相关系数指标,对室内16个不同位置的污染源展开位置辨识.结果最终有15个污染源的位置辨识结果准确;辨识效果与辨识时段的选取有关;在0～200 s的统计区间内,整体辨识效果呈现先升高,后略微下降并趋于稳定的趋势;辨识效果最好的时段出现在τ=80～120 s,此段时间内的平均辨识得分率高达94. 375%;在位置辨识准确的前提下,通过计算线性回归方程的斜率完成了对污染源散发强度的辨识,辨识误差保持在±0. 005 g/s的可控范围内.结论 Pearson评价法可以实现对污染源的位置和散发强度的有效辨识.

基于Pearson相关检验的EMD-ELM瓦斯涌出量时序预测研究

为分析地质条件和开采条件等相关因素对煤矿瓦斯涌出量时间序列的影响,挖掘小样本条件下多源监测信息与呈非线性特征的瓦斯涌出量之间的关系,实现瓦斯涌出量的精准预测,在时间序列理论基础上,选用经验模态分解(EMD)法和极限学习机(ELM)构建基于监测信息的瓦斯涌出量EMD-ELM多源时变预测模型。对于小样本监测数据,采用Pearson相关检验选定主要瓦斯涌出量影响因素,通过EMD深层分解瓦斯涌出量监测序列,求出瓦斯涌出量本质模函数(IMF),提取瓦斯涌出量非线性时间序列的样本特征,以及对应瓦斯涌出量序列的高频波动项和低频趋势项;对各项分量分别建立极限学习机预测模型,将分量预测结果等权求和作为瓦斯涌出量的最终预测值。以某煤矿瓦斯涌出量监测样本作为检验数据,检验结果显示经验模态分解能有效简化数据,模型均方误差为0.017,模型精度和泛化能力相比直接预测有所提高。

基于Pearson相关系数及线性回归的新型城镇化与耕地面积变化研究——以宿迁市为例

根据新型城镇化的内涵,利用复合指标法从人口城镇化、经济发展、社会公共环境三个层面构建宿迁市新型城镇化综合评价指标体系。在此基础上,利用Pearson相关性分析和回归分析的方法对宿迁市新型城镇化水平与耕地面积之间的关系进行分析。结果显示,2002-2012年耕地面积变化与新型城镇化发展成负相关,到2012-2016年趋势变化转为正相关,可以得出耕地面积变化与城镇化发展不是完全对立的,并由此提出相关建议。

Natural variations and geographical distributions of seed carotenoids and chlorophylls in 1167 Chinese soybean accessions

Understanding the composition and contents of carotenoids in various soybean seed accessions is important for their nutritional assessment.This study investigated the variability in the concentrations of carotenoids and chlorophylls and revealed their associations with other nutritional quality traits in a genetically diverse set of Chinese soybean accessions comprised of cultivars and landraces.Genotype,planting year,accession type,seed cotyledon color,and ecoregion of origin significantly influenced the accumulation of carotenoids and chlorophylls.The mean total carotenoid content was in the range of 8.15–14.72μg g–1 across the ecoregions.The total carotenoid content was 1.2-fold higher in the landraces than in the cultivars.Soybeans with green cotyledons had higher contents of carotenoids and chlorophylls than those with yellow cotyledons.Remarkably,lutein was the most abundant carotenoid in all the germplasms,ranging from 1.35–37.44μg g–1.Carotenoids and chlorophylls showed significant correlations with other quality traits,which will help to set breeding strategies for enhancing soybean carotenoids without affecting the other components.Collectively,our results demonstrate that carotenoids are adequately accumulated in soybean seeds,however,they are strongly influenced by genetic factors,accession type,and germplasm origin.We identified novel germplasms with the highest total carotenoid contents across the various ecoregions of China that could serve as the genetic materials for soybean carotenoid breeding programs,and thereby as the raw materials for food sectors,pharmaceuticals,and the cosmetic industry.

基于MLP-BiLSTM-TCN组合的超短期风电功率预测

风电功率受气象等因素影响具有很强的随机性和波动性,传统单一神经网络模型预测精度较低,难以满足需求。基于此,提出了多层感知机(multi-layer perceptron,MLP)-双向长短期记忆(bidirectional long short-term memory,BiLSTM)-时间卷积网络(temporal convolutional network,TCN)组合的超短期风电功率预测模型。首先使用皮尔逊相关系数处理数据,将相关性高的特征和每个预测点前6个真实值作为输入,然后生成3种基础神经网络模型MLP、BiLSTM、TCN,并根据隐藏单元设置不同参数,每个模型选取4个参数结果即组成12种基础模型。最后采用线性规划(linear programming,LP)法求出不同权值,以达到结果的最优化。以新疆某风电场为例对所提组合模型进行验证,绝对平均绝对误差(mean absolute error,MAE)为2.975 MW,均方根误差(root mean square error,RMSE)为5.474 MW,相较于其他模型,所提组合模型具有更高的预测精度。该研究能在风电功率预测中进行实时调度,解决电网调频问题,对组合模型预测研究有促进作用。

基于GST-Hough变换的LFM信号识别方法

针对线性调频(linear frequency modulated,LFM)信号在低信噪比条件下的信号检测问题,提出将广义S变换(generalized S transform,GST)与Hough变换相结合(generalized S transform based on Hough transform,GSTH)信号检测方法。从理论层面推导出LFM信号在进行GST后对应的参数特性,论证Hough变换的可行性,推导出GSTH变换后LFM信号与噪声的概率密度分布函数,给出了基于奈曼-皮尔逊准则进行峰值检测时,检测门限的计算方法与确定流程。利用GST时频聚焦性提供良好的直线线性,有易于Hough变换的直线检测,提升变换后主峰峰值并降低副峰高度。通过与WHT(Wigner-Hough transform)、分数阶傅里叶变换与周期WHT算法的仿真对比,定量评估算法的适用性,并与经典算法对比,定性的描述出算法良好的时频聚焦性,凸显GSTH算法在强噪声背景下具有更好的检测精度与适用范围。

利用相关分析法对S型海洋重力仪数据进行分析与改正

在理想情况下,海洋重力仪的测量数据经厄特弗斯校正和交叉耦合改正后应不受载体运动状态的影响,即与仪器的运动状态没有任何的相关性。但实际测量结果表明,当海况较差时,载体的运动状态对S型海洋重力仪的测量数据有着较为明显的影响。通过对实测重力数据和仪器运动状态的相关分析,指出仪器原来交叉耦合改正的不足,并且利用线性回归进一步对重力数据进行了改正。

基于一种DES算法改进体制的研究

本文用改进的S -盒替换原来的S -盒 ,用内部CBC三重加密模式以及相关密钥S -盒等方式构造出一种改进的DES加密体制。这种体制增强了线性密码分析攻击、差分密码分析攻击、穷举攻击和选择明文攻击的难度 ,而它加解密的速度并没有受到影响 ,相反通过使用相关密钥S -盒硬件芯片 。

基于多元线性回归的空气质量指数预测模型

随着中国社会的发展和人口数量的增多,空气污染环境问题表现突出,对人们的生活环境以及生态系统造成严重破坏。本文基于《中国统计年鉴2022》的相关数据,选取PM_(10)、O_(3)、SO_(2)、PM_(2.5)、N0_(2)、CO六个影响空气质量指数的指标。首先,通过相关性分析模型验证六个影响指标都对空气质量指数有较强的相关性;其次,通过各污染物浓度计算相应污染物分指数,利用多元线性回归模型建立空气质量指数(air quality index, AQI)与各影响指标之间的数学模型,利用该模型对城市空气质量指数进行预测;最后,将真实值与预测值使用皮尔逊相关性检验对模型评估,准确率达到0.928,能够反映空气质量的总体状况。

义昌江流域水质水量演变规律分析

研究义昌江流域降水量、径流量和水质指标演变规律,可为流域水资源开发利用、水环境保护和水生态改善提供数据支撑。运用线性趋势回归检验法研究其变化趋势,采用有序聚类分析法对其进行突变分析;分析水质指标时空变化,并进行水质评价;基于Pearson相关分析和双累积曲线法,研究义昌江流域水质变化的原因。结果表明:降水量、径流量、溶解氧、高锰酸盐指数、氨氮、粪大肠菌群变化趋势不明显。一般情况下,下游大冲口断面水质比上游断面好;从季节变化来看,冬季水质大部分合格、水质较好;从年变化来看,水质变化波动;径流对溶解氧、氨氮的影响比降水和水温更大,水温对高锰酸盐指数、粪大肠菌群的影响比降水和径流更大;水质指标变化受自然因素和人类活动共同影响。

反相高效液相色谱中同系物组分的线性S－logk＇w相关分析

研究了反相高效液相色谱中不同冲洗剂体系中同系物组分的线性Ｓ－ｌｏｇｋ＇ｗ相关分析（ＬＳＬＣ）。ＬＳＬＣ分析可以用来表征反相高效液相色谱的柱系统，在ＬＳＬＣ方程Ｓ＝ｐ·ｌｏｇｋ＇ｗ＋ｑ中，ｑ／ｐ主要表征了柱相比的对数（ｌｏｇΦ），而且对给定冲洗剂系统是一个常数；由ＬＳＬＣ所获得的相比并不显著的随同系物的种类的变化而变化，它是随Ｇ＿１８键含量的增加而增加的。

P值大小不等价于差异或相关性大小

目的探讨P值与差异或相关性大小的关系。方法固定样本统计量[均数和标准差、样本率、相关系数、OR(95%CI)],分别模拟独立样本t检验、两组样本率比较的卡方检验、Pearson相关分析、单因素Logistic回归分析的数据,观察不同样本量下P值差异。结果两组样本统计量相同或单组样本统计量一致但样本量不同时,P值可能大于0.05、介于0.001~0.05或小于0.001。结论P值受样本量影响大,其大小不等价于差异或相关性大小。

基于s-LTP和相似度匹配的人脸识别算法

针对传统的LTP算子计算过程复杂、不能反映图像局部与整体的关系的不足,提出了一种带符号局部三值模式(s-LTP)算子,首先利用2,1,0三值编码计算出像素块的LTP特征值,再通过比较该像素块中心点像素值与图像平均像素值的大小对LTP特征值赋正负号,得到s-LTP特征值。然后提出了一种基于皮尔逊相关系数的相似度匹配方法,通过大量实验得到最佳皮尔逊相关系数阈值和相似度阈值,最后通过计算待测样本图像和模板图像的相似度对待测样本图像进行分类识别。在Jaffe、ORL、Yale和一个自建的人脸数据库上进行实验,和其他人脸识别方法相比发现该算法识别率更高、所需识别时间更少。

基于SPOT-5光谱和纹理信息的湘西森林生态功能指数遥感预测模型构建

以湘西区域SPOT-5遥感影像为基础,提取样地纹理和光谱信息,以一类调查数据的生态指数为因变量,所对应的纹理和光谱指数为自变量,采用全子集回归来构建预估模型,并采用留一交叉验证进行模型检验。结果表明,构建模型的判定系数R2adj为0.5071,留一交叉验证结果R2cv=0.4860,模型的残差呈均匀的分布在0附近,没有明显的变化趋势。此外,SW检验和NCV检验结果显示残差的正态性和等方差性,表明构建的森林生态功能指数遥感预估模型能够预估森林生态功能指数,为生态功能的快速、经济和定量的评价提供数据支持,为有效森林管理以及决策的制定提供理论支持。