Meta-modeling using generalized regression neural network and particle swarm optimization

Swarm intelligence (SI) and evolutionary computation (EC) algorithms are often used to solve various optimization problems. SI and EC algorithms generally require a large number of fitness function evaluations (i.e., higher computational requirements) to obtain quality solutions. This requirement becomes more challenging when optimization problems are associated with computationally expensive analyses and/or simulation tasks. To tackle this issue, meta-modeling has shown successful results in improving computational efficiency by approximating the fitness or constraint functions of these complex optimization problems. Meta-modeling approaches typically use polynomial regression, kriging, radial basis function network, and support vector machines. Less attention has been given to the generalized regression neural network approach, and yet, it offers several advantages. Specifically, the model construction process does not require iterations. Its only one parameter is known to be less sensitive and usually requires less effort in selecting an optimal parameter. We use generalized regression neural network in this paper to construct meta-models and to approximate the fitness function in particle swarm optimization. To assess the performance and quality of these solutions, the proposed meta-modeling approach is tested on ten benchmark functions. The results are promising in terms of the solution quality and computational efficiency, especially when compared against the results of particle swarm optimization without meta-modeling and several other meta-modeling methods in previously published literature.     

A recommendation system for meta-modeling: A meta-learning based approach

Various meta-modeling techniques have been developed to replace computationally expensive simulation models. The performance of these meta-modeling techniques on different models is varied which makes existing model selection/recommendation approaches (e.g., trial-and-error, ensemble) problematic. To address these research gaps, we propose a general meta-modeling recommendation system using meta-learning which can automate the meta-modeling recommendation process by intelligently adapting the learning bias to problem characterizations. The proposed intelligent recommendation system includes four modules: (1) problem module, (2) meta-feature module which includes a comprehensive set of meta-features to characterize the geometrical properties of problems, (3) meta-learner module which compares the performance of instance-based and model-based learning approaches for optimal framework design, and (4) performance evaluation module which introduces two criteria, Spearman's ranking correlation coefficient and hit ratio, to evaluate the system on the accuracy of model ranking prediction and the precision of the best model recommendation, respectively. To further improve the performance of meta-learning for meta-modeling recommendation, different types of feature reduction techniques, including singular value decomposition, stepwise regression and ReliefF, are studied. Experiments show that our proposed framework is able to achieve 94% correlation on model rankings, and a 91% hit ratio on best model recommendation. Moreover, the computational cost of meta-modeling recommendation is significantly reduced from an order of minutes to seconds compared to traditional trial-and-error and ensemble process. The proposed framework can significantly advance the research in meta-modeling recommendation, and can be applied for data-driven system modeling.     

Genetic programming through bi-objective genetic algorithms with a study of a simulated moving bed process involving multiple objectives

A new bi-objective genetic programming (BioGP) technique has been developed for meta-modeling and applied in a chromatographic separation process using a simulated moving bed (SMB) process. The BioGP technique initially minimizes training error through a single objective optimization procedure and then a trade-off between complexity and accuracy is worked out through a genetic algorithm based bi-objective optimization strategy. A benefit of the BioGP approach is that an expert user or a decision maker (DM) can flexibly select the mathematical operations involved to construct a meta-model of desired complexity or accuracy. It is also designed to combat bloat – a perennial problem in genetic programming along with over fitting and under fitting problems. In this study the meta-models constructed for SMB reactors were compared with those obtained from an evolutionary neural network (EvoNN) developed earlier and also with a polynomial regression model. Both BioGP and EvoNN were compared for subsequent constrained bi-objective optimization studies for the SMB reactor involving four objectives. The results were also compared with the previous work in the literature. The BioGP technique produced acceptable results and is now ready for data-driven modeling and optimization studies at large.     

A Meta-Modeling Approach for Autonomous Driving Scenario Based on STTD

In the current autonomous driving scenario modeling and simulation field, autonomous driving modeling driven by Spatio-Temporal Trajectory Data (STTD) is a key problem, which is significant to improve the safety of the system. In recent years, great progress has been achieved in the modeling and application of STTD, and the application of this data in specific fields has attracted wide attention. However, because STTD has diversity and complexity as well as massive, heterogeneous, dynamic characteristics, the research in the safety-critical field modeling still faces challenges, including unified metadata of spatio-temporal trajectories, meta-modeling methods based on STTD, data processing based on the data analysis of spatio-temporal trajectories, and data quality evaluation. In view of the modeling requirements in the field of autonomous driving, a meta-modeling approach is proposed to construct spatio-temporal trajectory metadata based on Meta Object Facility (MOF) meta-modeling system. According to the characteristics of spatio-temporal trajectory data and autonomous driving domain knowledge, a meta-model of spatio-temporal trajectory data is constructed. Then, we study the modeling approach of autonomous driving safety-critical scenarios based on the spatio-temporal trajectory data meta-modeling technology system, use the modeling language ADSML for automatic instantiation of safety-critical scenarios, and construct a library of safety-critical scenarios, aiming to provide a feasible approach for the modeling of such safety-critical scenarios. Combined with the scenarios of lane changing and overtaking, the effectiveness of the meta-modeling method for autonomous driving safety scenarios driven by spatio-temporal trajectory data is demonstrated, which lays a solid foundation for the construction, simulation, and analysis of the model.     

Advanced topic modeling for social business intelligence

Social business intelligence combines corporate data with user-generated content (UGC) to make decision-makers aware of the trends perceived from the environment. A key role in the analysis of textual UGC is played by topics, meant as specific concepts of interest within a subject area. To enable aggregations of topics at different levels, a topic hierarchy has to be defined. Some attempts have been made to address the peculiarities of topic hierarchies, but no comprehensive solution has been found so far. The approach we propose to model topic hierarchies in ROLAP systems is called meta-stars. Its basic idea is to use meta-modeling coupled with navigation tables and with dimension tables: navigation tables support hierarchy instances with different lengths and with non-leaf facts, and allow different roll-up semantics to be explicitly annotated; meta-modeling enables hierarchy heterogeneity and dynamics to be accommodated; dimension tables are easily integrated with standard business hierarchies. After outlining a reference architecture for social business intelligence and describing the meta-star approach, we formalize its querying expressiveness and give a cost model for the main query execution plans. Then, we evaluate meta-stars by presenting experimental results for query performances and disk space.     

Optimization of self-organizing polynomial neural networks

The main disadvantage of self-organizing polynomial neural networks (SOPNN) automatically structured and trained by the group method of data handling (GMDH) algorithm is a partial optimization of model weights as the GMDH algorithm optimizes only the weights of the topmost (output) node. In order to estimate to what extent the approximation accuracy of the obtained model can be improved the particle swarm optimization (PSO) has been used for the optimization of weights of all node-polynomials. Since the PSO is generally computationally expensive and time consuming a more efficient Levenberg–Marquardt (LM) algorithm is adapted for the optimization of the SOPNN. After it has been optimized by the LM algorithm the SOPNN outperformed the corresponding models based on artificial neural networks (ANN) and support vector method (SVM). The research is based on the meta-modeling of the thermodynamic effects in fluid flow measurements with time-constraints. The outstanding characteristics of the optimized SOPNN models are also demonstrated in learning the recurrence relations of multiple superimposed oscillations (MSO).     

基于PSO优选参数的SVR水质参数遥感反演模型

为进一步提高多光谱图像水质反演的精度,提出了一种基于PSO优选参数的SVR水质参数遥感反演模型.该模型利用高分辨率多光谱遥感SPOT-5数据和水质实地监测数据,采用交叉验证CV(cross validation)估计模型推广误差并使用PSO优选SVR模型参数,实现了模型参数的自动全局优选,在训练好的SVR模型基础之上对水质进行反演.以渭河陕西段为例进行实证研究,实验结果表明,本文提出的水质反演模型较常规的线性回归模型有更高的反演精度,为内陆河流环境遥感监测提供了一种新方法.     

函数式程序模板元编程的元建模实现方法

针对函数式程序模板元编程的通用性问题,以应用类型系统ATS（Applied Type System）为例,提出了一种基于元建模的模板元编程实现方法。基于ATS模板元编程给出从枚举类型Datatype到Function的生成实例;通过元建模构造了包含Datatype与Function定义的ATS元模型;详细描述了Datatype模型到Function模型的转换;最后以一个基于元建模的ATS模板元编程为例,讨论了该方法的使用效果。实验结果表明该方法可以提高ATS模板元编程的通用性。     

基于改进的PSO-SVM室内走廊环境WSN信道建模

随着无线传感器在室内的应用越来越普及,无线传感器网络(WSN)室内信道模型的研究成为了通信行业的一个热点。针对室内走廊环境进行了无线传感器网络的信号传输衰落特性测试,并根据测试所得数据利用粒子群算法优化支持向量机(PSO-SVM)进行了数据回归分析,通过选取交叉验证折数来提高精度,建立了PSO-SVM无线电波路径损耗模型。模型与双折线对数距离路径损耗模型进行对比分析,实验结果表明PSO-SVM模型预测精度优于双折线对数距离路径损耗模型,能很好的表征室内走廊环境下的WSN电波衰落特性。     

可复用MOF元建模框架的研究

在MDA场景下,元模型是实现平台无关模型和平台相关模型转换的核心.提出通过元层模型和模型层模型的Down-Up机制给出可复用的MOF元建模框架.元建模框架由MOF BootStrap模型自举,并且内置MOF Model,此框架可在任意多层元建模中复用.此外,给出了模型工程模型和模型迁移剪枝算法,并提出了采用模型工程统一对象空间实现模型实例复用的方法.此框架的研究对于指导具体建模工具的实现有重要意义.     

考虑随机不确定性的常态仿真模型确认试验设计

模型确认试验是一种新的试验,其目的在于度量仿真模型的可信度.为了得到低成本、高可信度的模型确认试验方案,提出一种随机不确定性模型确认试验设计方法.首先,基于面积确认度量指标提出一种新的无量纲的模型确认度量指标(面积确认度量指标因子),并且在其基础上发展了基于专家系统的仿真模型准确性定性评判准则;然后,建立随机不确定性模型确认试验优化设计模型,提出该优化模型的求解方法;最后,通过两个数值算例对提出的模型确认试验设计方法进行验证.结果表明,小样本情况下,试验方案的随机性会影响模型评判结果的可信度;面积度量指标因子随试验样本数量的增加而收敛;随机不确定性模型确认试验设计方法能够避免试验方案对模型确认结果的影响.     

fMRI脑图的感知状态分析——回归模型及其寻优的非同质检验

脑功能核磁共振图像fMRI的特点是定位准确,但信噪比低、数据量大。对fMRI数据的泛回归模型的超参数寻优问题作了分析,提出基于非同质检验的超参数确认方法,重点比较了它在线性和非线性的回归方式（包括岭回归,支持向量回归,Elman递归神经网络）下针对不同外界环境特征的回归能力差异,实验所采用原始数据均来自PBAIC2006,结果表明,该方法在对相关领域知识较少依赖的前提下,具有较好的稳定性和泛化能力;同时在所涉及到的回归方法当中,线性方法的实现简单、有效,在计算代价上低于其他方法,对多种外界特征具有较高的预测能力。     

Solving Local Cost Estimation Problem for Global Query Optimization in Multidatabase Systems

To meet users' growing needs for accessing pre-existing heterogeneous databases, a multidatabase system (MDBS) integrating multiple databases has attracted many researchers recently. A key feature of an MDBS is local autonomy. For a query retrieving data from multiple databases, global query optimization should be performed to achieve good system performance. There are a number of new challenges for global query optimization in an MDBS. Among them, a major one is that some local optimization information, such as local cost parameters, may not be available at the global level because of local autonomy. It creates difficulties for finding a good decomposition of a global query during query optimization. To tackle this challenge, a new query sampling method is proposed in this paper. The idea is to group component queries into homogeneous classes, draw a sample of queries from each class, and use observed costs of sample queries to derive a cost formula for each class by multiple regression. The derived formulas can be used to estimate the cost of a query during query optimization. The relevant issues, such as query classification rules, sampling procedures, and cost model development and validation, are explored in this paper. To verify the feasibility of the method, experiments were conducted on three commercial database management systems supported in an MDBS. Experimental results demonstrate that the proposed method is quite promising in estimating local cost parameters in an MDBS.     

时空轨迹数据驱动的自动驾驶场景元建模方法

时空轨迹数据驱动的汽车自动驾驶场景建模,是当前汽车自动驾驶领域中驾驶场景建模、仿真所面临的关键问题,对于提高系统的安全性具有重要的研究意义.近年来,随着时空轨迹数据建模及应用研究的快速发展,时空轨迹数据应用于特定领域建模的研究引起人们的广泛关注.但是,由于时空轨迹数据所反映的现实世界的多元性和复杂性以及时空轨迹数据的海...     

基于回归模型的数据挖掘研究

回归分析是数据挖掘系统中的重要方法之一,本文主要研究如何利用回归模型来进行数据挖掘建模,介绍回归模型的3种类型,基于最小二乘法的参数估计和方程的显著性校验,并提出模型的优化方案,包括离群点的检验处理,模型形式的改进和回归自变量的选取。最后根据实例分析其在数据挖掘中的应用。     

HyperMoVal: Interactive Visual Validation of Regression Models for Real‐Time Simulation

During the development of car engines, regression models that are based on machine learning techniques are increasingly important for tasks which require a prediction of results in real‐time. While the validation of a model is a key part of its identification process, existing computation‐ or visualization‐based techniques do not adequately support all aspects of model validation. The main contribution of this paper is an interactive approach called HyperMoVal that is designed to support multiple tasks related to model validation: 1) comparing known and predicted results, 2) analyzing regions with a bad fit, 3) assessing the physical plausibility of models also outside regions covered by validation data, and 4) comparing multiple models. The key idea is to visually relate one or more n‐dimensional scalar functions to known validation data within a combined visualization. HyperMoVal lays out multiple 2D and 3D sub‐projections of the n‐dimensional function space around a focal point. We describe how linking HyperMoVal to other views further extends the possibilities for model validation. Based on this integration, we discuss steps towards supporting the entire workflow of identifying regression models. An evaluation illustrates a typical workflow in the application context of car‐engine design and reports general feedback of domain experts and users of our approach. These results indicate that our approach significantly accelerates the identification of regression models and increases the confidence in the overall engineering process.     

Multi-objective optimization design for a sand crab-inspired compliant microgripper

A large working travel and a minimal stress are the most critical characteristics of a microgripper but they are conflicted each other. This paper develops a new efficient hybrid algorithm to solve the multi-objective optimization design for a sand bubbler crab-inspired compliant microgripper. The structure of sand bubbler crab-inspired compliant microgripper is inspired from the profile of sand bubbler crab. A surrogate-assisted multi-objective optimization is conducted by developing a hybrid approach of finite element analysis, response surface method, Kigring metamodel and multi-objective genetic algorithm. First, the data are collected by integrating the finite element analysis and response surface method. Subsequently, in the types of common surrogates, Kigring metamodel is adopted as an efficient tool to approximate the objective functions. And then, the Pareto-optimal fronts are found via the multi-objective genetic algorithm. The results indicated that the optimal results are at the displacement of 5999.9 µm and stress of 330.68 MPa. The results revealed that the optimized results are highly consistent with both the validation results. The accuracy of the surrogate models showed that the regression model is a good prediction. The proposed approach is useful tool to solve complex optimization designs.

     

基于核偏最小二乘的锌层重量预测模型

为了给带钢热镀锌生产的质量控制提供必要的决策支持和分析手段,针对气刀对锌层重量的控制工艺,提出了基于核偏最小二乘回归的锌层重量预测模型。利用核函数将低维空间的非线性回归转化为高维空间的线性回归,克服了实际生产工艺中非线性因素对预测模型的不利影响。应用鞍山钢铁集团公司带钢热镀锌的生产实际数据进行验证,结果表明,基于核偏最小二乘的锌层重量预测方法与线性偏最小二乘、BP神经网络等方法相比,具有更好的预测精度。     

铁水硅含量的混沌粒子群支持向量机预报方法

提出一种基于混沌粒子群优化(CPSO)的支持向量回归机(SVR)参数优化算法, 并使用该算法建立高炉铁水硅含量预测模型(CPSO–SVR), 对某大型钢铁厂高炉铁水硅含量的实际采集数据进行预测, 结果表明基于混沌粒子群优化算法寻优的参数建立的铁水硅含量支持向量回归预测模型具有良好的预测效果. 与最小二乘支持向量回归机(LS–SVR)、使用粒子群优化算法训练的神经网络(PSO–NN)进行比较, CPSO–SVR模型对铁水硅含量进行预测时预测绝对误差小于0.03的样本数占总测试样本数的百分比达到90%以上, 预测效果明显优于PSO–NN, 且比LS–SVR稳定性更强, 可用于高炉铁水硅含量的实际预测, 表明混沌粒子群优化算法是选取SVR参数的有效方法.     

基于概率支持向量回归的产品设计时间预测模型

为使产品设计时间预测既克服小样本、异方差噪声问题,又提供除预测值以外的其他有用信息,建立概率支持向量回归(PSVR)模型。首先,在异方差回归模型基础上设计概率约束条件,以使预测值以较大概率位于真实值的某邻域,结合具有参数不敏感损失函数的支持向量回归确定优化目标,提出PSVR。然后,将最大完工时间知识嵌入进PSVR的约束条件,用以确定真实值邻域的宽度,将交叉验证与遗传算法相结合以确定PSVR的相关参数。最后,以注塑模具设计的实例进行分析,结果表明基于PSVR的时间预测方法可同时提供有效的预测值和预测区间。