On the prediction of solubility of alkane in carbon dioxide using the LSSVM algorithm

The increasing global energy demand and declination of oil reservoir in recent years cause the researchers attention focus on the enhancement of oil recovery approaches. One of the extensive applicable methods for enhancement of oil recovery, which has great efficiency and environmental benefits, is carbon dioxide injection. The CO₂ injection has various effects on the reservoir fluid, which causes enhancement of recovery. One of these effects is extraction of lighter components of crude oil, which straightly depends on solubility of hydrocarbons in carbon dioxide. In order to better understand of this parameter, in this study, Least squares support vector machine (LSSVM) algorithm was developed as a novel predictive tool to estimate solubility of alkane in CO₂ as function of carbon number of alkane, carbon dioxide density, pressure, and temperature. The predicting model outputs were compared with the extracted experimental solubility from literature statistically and graphically. The comparison showed the great ability and high accuracy of developed model in prediction of solubility.     

基于中文自然语言的纹理分类新方法

提出了一种新的基于中文自然语言纹理描述词的纹理分类方法,建立了自然纹理分类体系,并用最小二乘支持向量机对纹理进行分类,实现了纹理的视觉特征到语义描述的转换.实验结果证明,该方法在图像理解和基于内容的图像检索中有助于缩小纹理特征的数学描述和人类理解之间的"语义鸿沟".     

基于遗传算法和LSSVM的网络安全事件发生频率预测

Since the frequency of network security incidents is nonlinear, traditional prediction methods such as ARMA, Gray systems are difficult to deal with the problem. When the size of sample is small, methods based on artificial neural network may not reach a high degree of preciseness. Least Squares Support Vector Machines (LSSVM) is a kind of machine learning methods based on the statistics learning theory, it can be applied to solve small sample and non-linear problems very well. This paper applied LSSVM to predict the occur frequency of network security incidents. To improve the accuracy, it used an improved genetic algorithm to optimize the parameters of LSSVM. Verified by real data sets, the improved genetic algorithm (IGA) converges faster than the simple genetic algorithm (SGA), and has a higher efficiency in the optimization procedure. Specially, the optimized LSSVM model worked very well on the prediction of frequency of network security incidents.     

基于鲁棒最小二乘支持向量机的齿轮磨损预测

为了降低包含噪声的现场齿轮磨损数据对最小二乘支持向量机 (least squares support vector machine, LSSVM) 模型稳健性的影响, 采用迭代鲁棒最小二乘支持向量机 (iteratively robust least squares support vector machine, IRLSSVM) 对齿轮磨损数据进行建模和预报.首先, 增加权函数迭代次数以保证建模过程的鲁棒性;然后, 将具有全局搜索的耦合模拟退火 (coupled simulated annealing, CSA) 与局部优化的单纯形法 (simplex method, SM) 相结合的方法用于优化IRLSSVM模型超参数, 进而采用鲁棒交叉验证作为CSA-SM算法拟合目标函数, 提高IRLSSVM模型超参数优化过程的鲁棒性;最后, 利用K727840ZW变速箱现场齿轮磨损数据进行了数值实验, 结果验证了所提出方法的有效性.     

Development of Reliable Models for Determination of Required Monoethanolamine (MEA) Circulation Rate in Amine Plants

In gas sweetening plants, one of the most sensitive operating parameters is amine circulation rate that must be carefully examined to yield the optimum design for each application. In this study, effort has been made to use computational intelligences for accurate estimating the Monoethanolamine (MEA) circulation rate in amine treating unit. In the first method, optimal topology for feed-forward type neural network, particularly multi-layer perceptron (MLP), has been obtained. In the second method, least square version of support vector machine (LSSVM) algorithm has been employed for the application of interest. Results of this communication demonstrate that the presented models are capable of predicting MEA circulation rate precisely. Since there is no need to define the topology of the LSSVM model in advance, application of this type of modeling is more preferable.     

基于分阶段的LSSVM发酵过程建模

发酵过程建模是研究微生物发酵的重要课题,基于模型可实现被测参量的软测量、系统的优化控制。鉴于引入混合核函数的最小二乘支持向量机在过程建模中具有优良表现,采用基于混合核函数的最小二乘支持向量机建模。但由于发酵过程周期较长,最小二乘支持向量机的全局模型预测精度难以保证,算法复杂度很高,因此提出一种分阶段建模方法。首先,选择表征阶段特性的辅助变量,利用模糊C均值聚类算法对样本数据聚类,将发酵过程分成不同的阶段,然后为各个阶段分别建立最优混合核最小二乘支持向量机局部模型,最后将局部模型合成构成过程的完整模型。将此方法应用于青霉素发酵过程和重组大肠杆菌发酵过程中,验证了该方法的有效性。     

一种新的近红外光谱信息区间选择方法

基于间隔策略的信息波长选择是近红外光谱分析中广泛应用的一种方法。针对传统算法忽略非线性因素的缺点,首次考虑将最小二乘支持向量机（least-squares support vector machine,LSSVM）方法应用于间隔选择策略,进而提出了一种新的波长选择方法iLSSVM（interval LSSVM）。该算法克服了传统间隔选择算法依赖于线性模型的缺陷,对存在较强非线性的光谱数据能够准确地选择最优信息区间,极大地减少建模变量并显著改善模型预测精度。应用两组业界标准的光谱数据来验证该算法的性能,并和传统方法进行了比较。实验结果表明,在两组数据集上该算法取得的标准预测偏差（root mean square error of prediction,RMSEP）分别比全谱PLS建模降低了20%和4%,比传统的间隔偏最小二乘算法（interval partial least-squares,iPLS）降低了28%和2%。     

灰色模糊LSSVM预测模型在锌净化除钴中的应用

针对锌净化除钴过程生产数据存在噪声和系统参数缓慢变化的问题,提出一种基于灰色模糊LSSVM的钴离子浓度预测模型。对样本数据进行灰色累加,削弱原始数据序列中的噪声,使数据规律性增强,灰色累加后数据作为LSSVM输入,提高模型抗干扰能力和预测能力;由于锌净化除钴工序的系统参数随时间发生变化,提出对不同时期的样本赋予不同的模糊加权值;利用改进PSO的全局优化能力和快速收敛性,优化LSSVM模型的惩罚因子和核函数参数,避免人为选择参数的盲目性。对硫酸锌溶液净化除钴过程生产数据的仿真结果表明,灰色模糊LSSVM预测值能很好地跟踪实际值的变化趋势,满足钴离子浓度预测要求。     

基于混合函数的KICA-LSSVM故障分类方法及应用

利用核独立成分分析(KICA)进行非线性特征提取,然后用最小二乘支持向量机建立故障分类模型。研究表明,不同核函数对模型的性能有很大影响。利用已有核函数构造混合核函数,提出基于混合核函数的KI-CA-LSSVM故障分类方法,并应用到某石化企业的润滑油生产过程。实验结果表明该方法具有很高的分类和泛化能力。     

基于变异CPSO算法的LSSVM蒸发过程软测量

在分析混沌粒子群优化算法(CPSO)和最小二乘支持向量机(SVM)理论基础上,以某氧化铝厂蒸发过程为对象,采用带有末位淘汰机制的混沌粒子群优化算法优化支持向量机的参数,建立了基于变异CPSO算法的LS-SVM的氧化铝蒸发过程软测量模型,并与PSO-LSSVM、LSSVM模型比较,研究表明,ICPSO-LSSVM模型预测准确,泛化性能好,且该模型预测结果中相对误差小于5%的样本达到92.5%,最大相对误差仅为8.1%,均方差MSE为0.05153,模型具有较高的精度,其现场实施结果表明基本可以实现出口浓度的实时在线预估。