Fuzzy classifier identification using decision tree and multiobjective evolutionary algorithms

This paper presents a hybrid method for identification of Pareto-optimal fuzzy classifiers (FCs). In contrast to many existing methods, the initial population for multiobjective evolutionary algorithms (MOEAs) is neither created randomly nor a priori knowledge is required. Instead, it is created by the proposed two-step initialization method. First, a decision tree (DT) created by C4.5 algorithm is transformed into an FC. Therefore, relevant variables are selected and initial partition of input space is performed. Then, the rest of the population is created by randomly replacing some parameters of the initial FC, such that, the initial population is widely spread. That improves the convergence of MOEAs into the correct Pareto front. The initial population is optimized by NSGA-II algorithm and a set of Pareto-optimal FCs representing the trade-off between accuracy and interpretability is obtained. The method does not require any a priori knowledge of the number of fuzzy sets, distribution of fuzzy sets or the number of relevant variables. They are all determined by it. Performance of the obtained FCs is validated by six benchmark data sets from the literature. The obtained results are compared to a recently published paper [H. Ishibuchi, Y. Nojima, Analysis of interpretability-accuracy tradeoff of fuzzy systems by multiobjective fuzzy genetics-based machine learning, International Journal of Approximate Reasoning 44 (1) (2007) 4–31] and the benefits of our method are clearly shown.     

A novel approach of initial vectors construction in iterative target transformation factor analysis

The construction of the favorable initial iterative vectors is the key to iterative target transformation factor analysis (ITTFA). A tentative approach to construct the better initial vectors, which is based on the chromatographic information provided by evolving factor analysis (EFA), is proposed. A region, which contains the peak position at maximum height, is determined. Several elements in the region of each initial vector, instead of one element, are initialized as 1. The elements out of the region are initialized as 0. So it is not necessary to determine the exact peak position at maximum height for the resolution of partly overlapping chromatographic profiles, which may avoid the divergence brought by determination of the peak position at maximum height. In addition, it may give acceptable resolution for the embedded peaks. It is applied to resolve 2D-simulated data and experimental liquor GC/MS data, the resolutions are reasonable and improved.     

一种新的适用于水声信道的常模类盲均衡算法

为了缓解常数模盲均衡算法收敛速度缓慢、稳态误差大的问题,考虑到初始化权值对CMA算法的重要影响,本文利用遗传算法对CMA算法进行了有效改进,引入了小样本重用的思想,给出了一种新的适用于水声信道的常模类盲均衡算法。计算机仿真研究证明,该算法不仅大大加快了收敛速度,而且有效地降低了稳态误差。     

Problem of proper decomposition and initialization of acoustic and entropy modes in a gas affected by the mass force

The relations connecting perturbations specific for acoustic and entropy modes in an accelerated fluid or in a fluid affected by constant mass force, are derived. They allow to decompose the total vector of perturbations and the overall energy into acoustic and non-acoustic parts uniquely at any instant. In order to do this, three quantities are required, for example total perturbations in entropy, pressure and velocity. The evaluations are made in regard to the content of acoustic and non-acoustic parts of the total energy excluding its kinetic part. In some cases, exact relations may be derived.     

Activation of a single-chamber solid oxide fuel cell by a simple catalyst-assisted in-situ process

Initialization is a critical processing step that has thus far limited the application of the single-chamber solid oxide fuel cell (SC-SOFC). In-situ initialization of a SC-SOFC with a nickel-based anode by methane–air mixtures was investigated. Porous Ru–CeO₂ was used as a catalyst layer over a Ni-ScSZ cermet anode. Catalytic testing demonstrated Ru–CeO₂ had high activity for methane oxidation. The Ru in the catalyst layer catalyzed the formation of syngas, which successfully reduced the nickel oxide to metallic nickel in the anode. Single cells with a La_0.8Sr_0.2MnO₃ (LSM) cathode, initialized by this in-situ reduction method, delivered peak power densities of 205 and 327 mW cm⁻² at 800 °C and 850 °C, respectively. Such performances were better than those of the cell without the Ru–CeO₂ catalyst layer that was initialized by an ex-situ reduction method were.     

Initialization in semidefinite programming via a self-dual skew-symmetric embedding

The formulation of interior point algorithms for semidefinite programming has become an active research area, following the success of the methods for large-scale linear programming. Many interior point methods for linear programming have now been extended to the more general semidefinite case, but the initialization problem remained unsolved.In this paper we show that the initialization strategy of embedding the problem in a self-dual skew-symmetric problem can also be extended to the semidefinite case. This method also provides a solution for the initialization of quadratic programs and it is applicable to more general convex problems with conic formulation.     

基于独立分量分析的高光谱图像压缩

在对原始数据进行虚拟维数估计的基础上,提出了一种基于最大距离端元提取+独立分量分析(Independent Component Analysis,ICA)的高光谱图像有损压缩方案.该方案首先应用最大距离端元抽取法提取高光谱数据各端元矢量,然后用快速独立分量分析生成独立分量,最后使用2维分层树集合分裂(Set Partitioning In Hierarchical Trees,SPIHT)算法对各独立分量图进行编码.计算机仿真结果证明,该算法在取得高压缩率的同时,能很好地保持数据的谱特征,是一种高效的三维数据压缩方法.     

Arbitrary initialization for Chan–Vese model

Three problems about initialization of Chan–Vese model impeding its applications of Chan–Vese model, are discussed in this paper first by various experiments. Then, a novel method to solve these badly initialization problems is proposed, which can be divided into three parts: a hypothesis that image boundaries do not contain any feature of the objects; an examination criterion to check whether the current contour is suitable for continuing or not; and a method to revise the unsuitable current contour into a new one. The most important advantage of our proposed method is that it realizes the arbitrary initialization which can release the manual work in essence. Extensive experiments on synthetic and real images demonstrate the advantages of the proposed method.     

Effects of including spatial information in clustering multivariate image data

Multivariate image data provide detailed information in variable and image space. Most traditional clustering methods are based on variable information only and ignore spatial information. A method based on both variable and spatial information could improve the results substantially.

In this review, we study the benefits and the pitfalls of including spatial information in chemometric clustering techniques. Spatial information is taken into account in initialization of clustering parameters, during cluster iterations by adjusting the similarity measure or at a post-processing step. We illustrate the effect of taking spatial information into account by a univariate synthetic data set and two real-world multivariate data sets. We show that methods that include neighboring pixel information in the clustering procedure improve the performance accuracy of the clustering in most cases. Homogeneous regions in the image are better recognized and the amount of noise is reduced by these methods.