Stability-based preference selection in affinity propagation

Recently, as one of the most popular exemplar-based clustering algorithms, affinity propagation has attracted a great amount of attention in various fields. The advantages of affinity propagation include the efficiency, insensitivity to cluster initialization and capability of finding clusters with less error. However, one shortcoming of the affinity propagation algorithm is that, the clustering results generated by affinity propagation strongly depend on the selection of exemplar preferences, which is a challenging model selection task. To tackle this problem, this paper investigates the clustering stability of affinity propagation for automatically selecting appropriate exemplar preferences. The basic idea is to define a novel stability measure for affinity propagation, based on which we can select exemplar preferences that generate the most stable clustering results. Consequently, the proposed approach is termed stability-based affinity propagation (SAP). Experimental results conducted on extensive real-world datasets have validated the effectiveness of the proposed SAP algorithm.     

A novel self-adaptive extreme learning machine based on affinity propagation for radial basis function neural network

In this paper, a novel self-adaptive extreme learning machine (ELM) based on affinity propagation (AP) is proposed to optimize the radial basis function neural network (RBFNN). As is well known, the parameters of original ELM which developed by G.-B. Huang are randomly determined. However, that cannot objectively obtain a set of optimal parameters of RBFNN trained by ELM algorithm for different realistic datasets. The AP algorithm can automatically produce a set of clustering centers for the different datasets. According to the results of AP, we can, respectively, get the cluster number and the radius value of each cluster. In that case, the above cluster number and radius value can be used to initialize the number and widths of hidden layer neurons in RBFNN and that is also the parameters of coefficient matrix H of ELM. This may successfully avoid the subjectivity prior knowledge and randomness of training RBFNN. Experimental results show that the method proposed in this thesis has a more powerful generalization capability than conventional ELM for an RBFNN.     

Learning high-dimensional correspondence via manifold learning and local approximation

The recent years have witnessed a surge of interests of learning high-dimensional correspondence, which is important for both machine learning and neural computation community. Manifold learning–based researches have been considered as one of the most promising directions. In this paper, by analyzing traditional methods, we summarized a new framework for high-dimensional correspondence learning. Within this framework, we also presented a new approach, Local Approximation Maximum Variance Unfolding. Compared with other machine learning–based methods, it could achieve higher accuracy. Besides, we also introduce how to use the proposed framework and methods in a concrete application, cross-system personalization (CSP). Promising experimental results on image alignment and CSP applications are proposed for demonstration.     

Wavelet twin support vector machine

Twin support vector machine (TWSVM) is a research hot spot in the field of machine learning in recent years. Although its performance is better than traditional support vector machine (SVM), the kernel selection problem still affects the performance of TWSVM directly. Wavelet analysis has the characteristics of multivariate interpolation and sparse change, and it is suitable for the analysis of local signals and the detection of transient signals. The wavelet kernel function based on wavelet analysis can approximate any nonlinear functions. Based on the wavelet kernel features and the kernel function selection problem, wavelet twin support vector machine (WTWSVM) is proposed by this paper. It introduces the wavelet kernel function into TWSVM to make the combination of wavelet analysis techniques and TWSVM come true. The experimental results indicate that WTWSVM is feasible, and it improves the classification accuracy and generalization ability of TWSVM significantly.     

Nonlinear multiobjective model-predictive control scheme for wastewater treatment process

A nonlinear multiobjective model-predictive control (NMMPC) scheme, consisting of self-organizing radial basis function (SORBF) neural network prediction and multiobjective gradient optimization, is proposed for wastewater treatment process (WWTP) in this paper. The proposed NMMPC comprises a SORBF neural network identifier and a multiple objectives controller via the multi-gradient method (MGM). The SORBF neural network with concurrent structure and parameter learning is developed as a model identifier for approximating on-line the states of WWTP. Then, this NMMPC optimizes the multiple objectives under different operating functions, where all the objectives are minimized simultaneously. The solution of optimal control is based on the MGM which can shorten the solution time. Moreover, the stability and control performance of the closed-loop control system are well studied. Numerical simulations reveal that the proposed control strategy gives satisfactory tracking and disturbance rejection performance for WWTP. Experimental results show the efficacy of the proposed method.     

A multi-document summarization system based on statistics and linguistic treatment

The massive quantity of data available today in the Internet has reached such a huge volume that it has become humanly unfeasible to efficiently sieve useful information from it. One solution to this problem is offered by using text summarization techniques. Text summarization, the process of automatically creating a shorter version of one or more text documents, is an important way of finding relevant information in large text libraries or in the Internet. This paper presents a multi-document summarization system that concisely extracts the main aspects of a set of documents, trying to avoid the typical problems of this type of summarization: information redundancy and diversity. Such a purpose is achieved through a new sentence clustering algorithm based on a graph model that makes use of statistic similarities and linguistic treatment. The DUC 2002 dataset was used to assess the performance of the proposed system, surpassing DUC competitors by a 50% margin of f-measure, in the best case.     

An improved global-best harmony search algorithm for faster optimization

In this paper, an improved global-best harmony search algorithm, named IGHS, is proposed. In the IGHS algorithm, initialization based on opposition-based learning for improving the solution quality of the initial harmony memory, a new improvisation scheme based on differential evolution for enhancing the local search ability, a modified random consideration based on artificial bee colony algorithm for reducing randomness of the global-best harmony search (GHS) algorithm, as well as two perturbation schemes for avoiding premature convergence, are integrated. In addition, two parameters of IGHS, harmony memory consideration rate and pitch adjusting rate, are dynamically updated based on a composite function composed of a linear time-varying function, a periodic function and a sign function in view of approximate periodicity of evolution in nature. Experimental results tested on twenty-eight benchmark functions indicate that IGHS is far better than basic harmony search (HS) algorithm and GHS. In further study, IGHS has also been compared with other eight well known metaheuristics. The results show that IGHS is better than or at least similar to those approaches on most of test functions.     

Automatic classification and prediction models for early Parkinson’s disease diagnosis from SPECT imaging

Early and accurate diagnosis of Parkinson’s disease (PD) is important for early management, proper prognostication and for initiating neuroprotective therapies once they become available. Recent neuroimaging techniques such as dopaminergic imaging using single photon emission computed tomography (SPECT) with ¹²³I-Ioflupane (DaTSCAN) have shown to detect even early stages of the disease. In this paper, we use the striatal binding ratio (SBR) values that are calculated from the ¹²³I-Ioflupane SPECT scans (as obtained from the Parkinson’s progression markers initiative (PPMI) database) for developing automatic classification and prediction/prognostic models for early PD. We used support vector machine (SVM) and logistic regression in the model building process. We observe that the SVM classifier with RBF kernel produced a high accuracy of more than 96% in classifying subjects into early PD and healthy normal; and the logistic model for estimating the risk of PD also produced high degree of fitting with statistical significance indicating its usefulness in PD risk estimation. Hence, we infer that such models have the potential to aid the clinicians in the PD diagnostic process.     

Face recognition with Symmetric Local Graph Structure (SLGS)

Face recognition demonstrates the significant progress in the research field of biometric and computer vision. The fact is due to the current systems perform well under relatively control environments but tend to suffer when the present of variation in pose, illumination, and facial expression. In this work, a novel approach for face recognition called Symmetric Local Graph Structure (SLGS) is presented based on the Local Graph Structure (LGS). Each pixel is represented with a graph structure of its neighbours’ pixels. The histograms of the SLGS were used for recognition by using the nearest neighbour classifiers that include Euclidean distance, correlation coefficient and chi-square distance measures. AT&T and Yale face databases were used to be experimented with the proposed method. Extensive experiments on the face database clearly showed the superiority of the proposed approach over Local Binary Pattern (LBP) and LGS. The proposed SLGS is robust to variation in term of facial expressions, facial details, and illumination. Due to good performance of SLGS, it is expected that SLGS has a potential for application implementation in computer vision.