Improved Affinity Propagation Clustering Research Articles

Undersea Water Objects Detection and Classification Using Optimized Region-Based Convolutional Neural Network

Underwater autonomous operation is becoming increasingly crucial to avoid the hazardous in the environment of high-pressure deep-sea due to the significance of underwater investigation. The most crucial piece of technology for underwater-based task is intelligent computer vision. In an underwater environment, underwater vision requires good image quality, and illumination with better classification of sea objects. This work presented a novel technique of Optimized Region-based Convolutional Neural Network (ORCNN). For this work, the Gaussian filter is used to remove the noise and enhance the image quality during pre-processing and the Improved Affinity Propagation Clustering (IAPC) model segments the objects. After that, the Region-based Convolutional Neural Network (RCNN) model classifies various objects such as urchins, seagrass, fishes, and rocks in which the RCNN parameters are tuned via a light spectrum optimizer algorithm (LSOA). can also input the segmentation prediction images and the labeled images into the discriminant convolutional network and improve the segmentation accuracy of underwater images by further enhancing the essential characteristics of learning data through the confrontation training of generators and discriminators. The experimental results demonstrate that higher performance is provided by the newly developed ORCNN -Net predictive model when compared to other comparative algorithms while considering the negative and positive metrics.

Quality-related fault monitoring for multi-phase batch process based on multiway weighted elastic network

A quality-related fault monitoring method of multi-phase batch process based on multiway weighted elastic network is proposed in this paper. Firstly, to make the phase division for batch process more accurately, an improved affinity propagation clustering algorithm is developed. Secondly, a multiway weighted elastic network model is developed in each phase. On the one hand, quality-related subspace and quality-unrelated subspace are constructed in each phase to achieve dual monitoring of process fault and quality anomalies. On the other hand, kernel density estimation is used to measure the contribution of each element in each subspace to the fault. According to the difference of the contribution of each element to the fault, different weight is assigned to enhance the fault features and eliminate irrelevant features such as noise. Finally, support vector data description is used to establish monitoring indexes in both quality-related subspace and quality-unrelated subspace. Compared with traditional methods, the superiority and effectiveness of the proposed method have been verified by monitoring the penicillin fermentation process and the hot strip mill process.

An Improved AP Clustering Algorithm Based Critical Nodes Identification for Distribution Network With High PV Penetration

The power flow of the distributed network (DN) with high penetration of distributed PV changes significantly, such as bidirectional power flow and larger deviation of volage magnitude, under the variation of PV power output, and the number and the position of critical nodes would change accordingly. Hence, a data-driven nodes clustering and critical node identification method is proposed in this paper. Since the relationship between different nodes are nonlinear, the autoencoder method is firstly applied to obtain the unified features of the nodes connected with different number of branches. Using the unified features, the Frechet distance is calculated to demonstrate the similarity between any two nodes. The improved affinity propagation (AP) clustering algorithm is also proposed to classify the nodes into different clusters, and the number the clusters and the critical nodes in different clusters could be obtained automatically. The electrical distance is considered in the improved AP clustering algorithm, so that nodes in different branches with similar features are avoided to be classified into one cluster. The simulations are carried out on the IEEE 123-bus system and an actual DN system, the effectiveness and efficiency of proposed critical node identification method are verified.

Research on Designing an Industrial Product-Service System with Uncertain Customer Demands

The industrial product-service system (iPSS) is a kind of system engineering methodology, integration scheme, and business model to realize service value by adding intangible services in the whole life cycle. However, the design of the system involves many difficulties such as uncertain customer demands, strong subjectivity of the experience design, and long debugging times. Methods for solving upper problems are therefore essential. This paper presents a design model that integrates an improved affinity propagation (AP) clustering algorithm, quality function development (QFD), and axiomatic design (AD). The entire process of designing an iPSS can be split into three steps. First, uncertain customer demands is determined and standardized. Second, the functions of the product-service system are investigated. Finally, the structures of the system are determined. This paper examines the example of the control service of an iPSS for a water heater tank capping press. An improved AP clustering algorithm is used to determine standardized customer demands, the proposed QFD, and an AD integration model to initially establish a mapping between the customer demands domain and the function domain and clarify the design focus. Next, a QFD- and AD-integrated model is constructed to establish a mapping between the function domain and the structure domain and optimize the control scheme through the quality of its risk prediction. Finally the paper verifies that the upper process and methods can guide the design process effectively in production applications.

Synchronized Measurement Technology Supported Online Generator Slow Coherency Identification and Adaptive Tracking

In an electric power system, slow coherency can be applied to identify groups of the generating units, the rotors of which are swinging together against each other at approximately the same oscillatory frequencies of inter-area modes. This serves as a prerequisite-step of several emergency control schemes to identify power system control areas and improve transient stability. In this paper, slow coherent generators are grouped based on the direction and the strength of electromechanical coupling between different generators. The proposed algorithm performs low-pass filtering of generator frequency measurements. It adaptively determines the minimal number of the measurements to be processed in an observation window, and performs data selectivity to prevent mixing of interfering coherency indices. Finally, it adaptively tracks grouping changes of slow coherent generators and determines a finite number of groups for an improved affinity propagation clustering. The proposed algorithm is implemented as an online MATLAB program and verified in real-time using RTDS power system simulator with the integration of actual synchronized measurement technology components as hardware-in-the-loop. The obtained results demonstrate the effectiveness of the proposed algorithm for robust and near real-time identification of grouping changes of slow coherent generators during the quasi-steady-state and electromechanical transient period following a disturbance.

Phase partition and identification based on a two-step method for batch process

The batch process is a batch-repeated production process, which shows a multiple modal switching within the batch. This makes it difficult to use a single-mode analysis method to achieve accurate modeling and fault diagnosis. Therefore, a novel two-step phase partition idea is proposed based on improved affinity propagation (AP) clustering and sub-phase similarity diminishing scan (PSDS) method. In order to capture the dynamic characteristics of the modes switching, the improved AP clustering is used for phase preliminary partition, in which an effective method that is more suitable for complex batch process is proposed to calculate the similarity. For sub-phases generated by the phase preliminary partition, the internal process of each sub-phase also varies obviously with the development of duration, so an innovative method PSDS is proposed to implement phase fine partition. Then each sub-phase scanned by the PSDS method is identified and divided into stable parts and transition parts, which further reflects the change trend within the sub-phase. For the outliers and misclassification points that may arise during the process of phase partition, the solutions are put forward, respectively. Thus, the partition results with different characteristics are modeled and monitored separately by using the method of principal component analysis (PCA). A practical application on batch process, aircraft steering engine platform fault diagnosis experiment, is given to conform the feasibility and performance of the proposed method.

An Affinity Propagation Clustering Method Using Hybrid Kernel Function With LLE

Cluster analysis is important in data mining and clustering algorithms and has gained much attention during the last decade. However, it is a challenge to extract significant features from high-dimensional data and to rapidly provide satisfactory clustering results. This paper presents a new affinity propagation (AP) clustering method based on a hybrid kernel function with locally linear embedding, called LLE-HKAP, for the classification of gene expression datasets and standard UCI datasets. First, the locally linear embedding algorithm is used to reduce the dimension of the original dataset. Then, a novel AP clustering method based on a similarity measure with the hybrid kernel function is proposed. In this method, a new global kernel is defined that has high generalization ability. Meanwhile, a hybrid kernel function that linearly combines the proposed global kernel and the Gaussian kernel is defined to further enhance the learning ability of the global kernel. Moreover, the novel hybrid kernel is introduced to define a similarity measure and construct a similarity matrix of the AP clustering. Finally, the improved AP clustering algorithm is implemented on eight public gene expression datasets and eight standard UCI datasets for comparison with other related algorithms. The experimental results validate that our proposed clustering algorithm is efficient in terms of clustering accuracy and outperforms the currently available approaches with which it is compared.

Adaptive unsupervised classification of polarimetric SAR images using the improved affinity propagation clustering algorithm

ABSTRACTA robust automatic classification system is critical for polarimetric synthetic aperture radar (POLSAR) terrain processing. In most of the conventional classification methods, the number of classes could not be calculated before classification. In this article, we present a new unsupervised classification algorithm with an adaptive number of classes for POLSAR data which is capable of estimating the class numbers automatically. The approach is mainly composed of three operations. First, region-based feature map combining the polarimetric statistical and spatial information is constructed based on the turbopixel method. This is followed by a clustering step performed through an improved affinity propagation clustering with Wishart distance. Finally, the result of the improved affinity propagation clustering is classified using Wishart classifier. The proposed approach takes the spatial information into consideration and makes good use of the inherent statistical characteristics of POLSAR data. The performance of the proposed classification approach on three real datasets is presented and analysed, and the experimental results show that the approach provides more accurate estimation under the condition of various numbers of classes compared with existing methods.

An Improved Affinity Propagation Clustering Algorithm Based on Entropy Weight Method and Principal Component Analysis

Traditional affinity propagation algorithm has inefficient results when conducting clustering analysis of high dimensional data because effect lead to difficult find the proper class structure .In view of this, the author proposes an improved algorithm on the basis of Entropy Weight Method and Principal Component Analysis (EWPCA-AP). EWPCA-AP algorithm empowers the sample data by Entropy Weight Method, eliminate data irrelevant attributes by Principal Component Analysis, and travel with neighbor clustering algorithm, realization of high-dimensional data clustering in low dimension space. The numerical result of simulation experiment shows that the new EWPCA-AP algorithm can effectively eliminate the redundancy and irrelevant attributes of data and improve the performance of clustering. In addition, the proposed algorithm is applied in the area of the economy in our country and the clustering result is consistent with the real one. This algorithm provides a new intelligent evaluation method for Chinese economy.

An improved affinity propagation clustering algorithm based on principal component analysis and variation coefficient

To solve the 'dimension disaster' problem associated with traditional affinity propagation (AP) algorithm when dealing with high-dimensional data clustering, in this paper we propose an improved AP algorithm named as Affinity Propagation with Coefficient of Variation and Principal Component Analysis (CVPCA-AP). Initially, a dimension reduction was performed on the original data with a focus on keeping most information. Then, the variation coefficient was used to assign a power for the attribute of dimension data. After, the improved AP algorithm was applied for clustering. The simulation results show that the proposed algorithm has superior clustering performance as compared to the traditional AP algorithm. In addition, the proposed algorithm was applied to the clustering analysis for the listed companies, and some satisfactory results were also obtained. This method is enriching studies on intelligence theory and providing a novel reference tool to assist governments in making scientifically sound economic decisions.

Research of Individual Neural Network Generation and Ensemble Algorithm Based on Quotient Space Granularity Clustering

The aim of this paper is to develop an individual neural network generation and ensemble algorithm based on quotient space granularity clustering. Firstly, we give the characteristics of the quotient space granularity and affinity propagation(AP) clustering. Secondly, we introduce the quotient space concept to the AP clustering analysis, which can find an optimal granularity from all possible granularities. Then using improved AP clustering algorithm to seek optimal results of sample clustering and using different individual neural network to learn different categories of samples so that the degree of difference between networks and the generalization ability of neural network ensemble(NNE) can be improved. Further, according to the degree of correlation between the input data and the sample category to adaptively adjust ensemble weights. The algorithm proposed here is not only a method of generating the individual neural networks, but also can adaptively adjust ensemble weights of individual neural network. Experiments show that our proposed method is validity and correctness.

On Expanded and Improved Affinity Propagation Clustering Algorithm

Facing to the shortcoming of Affinity Propagation algorithm (AP), we present two expanded and improved AP algorithms. In the two algorithms, the AP algorithm based on Grid Cell (APGC) is an effective extension of AP algorithm on the level of grid cells, and the AP clustering algorithm based on Near neighbour Sampling (APNS) is trying to make some improving in time and space complexity. From some simulated comparison experiments of three algorithms, we know that APGC and APNS algorithms have evident improving than AP algorithm in time and space complexity. They can not only get a good clustering quality for massive data sets, but also filtrate noises and isolates well. So we can say they are two effective clustering algorithms with much applied prospect. At last, several research directions are presented.