Spectral-Locational-Spatial Manifold Learning for Hyperspectral Images Dimensionality Reduction

Dimensionality reduction (DR) plays an important role in hyperspectral image (HSI) classification. Unsupervised DR (uDR) is more practical due to the difficulty of obtaining class labels and their scarcity for HSIs. However, many existing uDR algorithms lack the comprehensive exploration of spectral-locational-spatial (SLS) information, which is of great significance for uDR in view of the complex intrinsic structure in HSIs. To address this issue, two uDR methods called SLS structure preserving projection (SLSSPP) and SLS reconstruction preserving embedding (SLSRPE) are proposed. Firstly, to facilitate the extraction of SLS information, a weighted spectral-locational (wSL) datum is generated to break the locality of spatial information extraction. Then, a new SLS distance (SLSD) excavating the SLS relationships among samples is designed to select effective SLS neighbors. In SLSSPP, a new uDR model that includes a SLS adjacency graph based on SLSD and a cluster centroid adjacency graph based on wSL data is proposed, which compresses intraclass samples and approximately separates interclass samples in an unsupervised manner. Meanwhile, in SLSRPE, for preserving the SLS relationship among target pixels and their nearest neighbors, a new SLS reconstruction weight was defined to obtain the more discriminative projection. Experimental results on the Indian Pines, Pavia University and Salinas datasets demonstrate that, through KNN and SVM classifiers with different classification conditions, the classification accuracies of SLSSPP and SLSRPE are approximately 4.88%, 4.15%, 2.51%, and 2.30%, 5.31%, 2.41% higher than that of the state-of-the-art DR algorithms.

A Spatial temporal classification analysis and visualization of tropical cyclone tracks in Bay of Bengal using GIS

Tropical cyclones (TC) are among the most devastating forms of natural hazards and the east coast of India is more prone to TC landfall causing significant socio-economic impacts. The Bay of Bengal (BoB) which forms the eastern sub basin of North Indian Ocean experiences the seasonally reversing monsoon, depression and TCs. In this study TC best track dataset of NIO basin over the period 1960–2016 from the IBTrACKs archive maintained by NOAA are used. In this work Firefly optimization is coupled with FCM for TC tracks classification. The classical FCM uses random initialization of cluster centroid often gets trapped in local optimal problem. The firefly algorithm is applied on the FCM for the cluster centroid computation, in this way improving the efficiency of FCM algorithm. The obtained classes are then projected in the visualization space. Visualizations are generated using the GIS environment to gain insight into the spatial distribution of TC tracks over decades. This study aims to develop a comprehensive assessment of variability in tropical cyclones with respect to ENSO modulated events, inter decadal variability and track sinuosity. In this paper we attempt to convey the cognitive results of comparative visualizations of TC tracks over Arabian Sea and Bay of Bengal sub basin during the strong, very strong El Niño and La Niña events. Finally we use Parallel Coordinate Plot (PCP) a visualization technique to demonstrate the correlation patterns of the TC parameters.

Segmentation of image based on k-means and modified subtractive clustering

Image segmentation has widespread applications in medical science, for example, classification of different tissues, identification of tumors, estimation of tumor size, surgery planning, and atlas matching. Clustering is a widely implemented unsupervised technique used for image segmentation mainly because of its simplicity and fast computation. However, the quality and efficiency of clustering-based segmentation is highly depended on the initial value of the cluster centroid. In this paper, a new hybrid segmentation approach based on k-means clustering and modified subtractive clustering is proposed. K-means clustering is a very efficient and powerful algorithm but it requires initialization of cluster centroid. And, the consistency of the clustering outcomes of k-means algorithm depends on the initial selection of the cluster center. To overcome this drawback, a modified subtractive clustering algorithm based on distance relations between cluster centers and data points is proposed which finds a more accurate cluster centers compared to the conventional subtractive clustering. These cluster centroids obtained from the modified subtractive clustering are used in k-means algorithm for segmentation of the image. The proposed method is compared with other existing conventional segmentation methods by using several synthetic and real images and experimental finding validates the superiority of the proposed method.

An Optimized DiscretizationApproach using k-Means Bat Algorithm

This study has proposed arelatively new discretization approach using k-means and Bat algorithm in preparation phase of classification problem. In essence, bat algorithm is applied to find the best search space solution. Eventually, the best search space solution is utilized to produce cluster centroid. The cluster centroid is very useful to determine appropriate breakpoint for discretization. The proposed discretization approach is applied in the experiments with continuous datasets. Decision Tree, k-Nearest Neighbours and Naïve Bayes classifiers are used in the experiments. The proposed discretization approach is evaluated against other existing approaches: K-Means algorithm, hybrid K-Means with Particle Swarm Optimization (PSO) and hybrid K-Means with Whale Optimization Algorithm (WOA).The classification performance is evaluated in terms of accuracy, recall, f-measure and receiver operating characteristic curve (ROC). To test the performance of the proposed algorithm, nine benchmark continuous datasets are used. The proposed algorithm show the better results compare to other approaches. The proposed algorithm performs better in discretization to solve classification problems.

A new approach for improving clustering algorithms performance

<p class="Abstract">Clustering represents one of the most popular and used Data Mining techniques due to its usefulness and the wide variations of the applications in real world. Defining the number of the clusters required is an application oriented context, this means that the number of clusters k is an input to the whole clustering process. The proposed approach represents a solution for estimating the optimum number of clusters. It is based on the use of iterative K-means clustering under three different criteria; centroids convergence, total distance between the objects and the cluster centroid and the number of migrated objects which can be used effectively to ensure better clustering accuracy and performance. A total of 20000 records available on the internet were used in the proposed approach to test the approach. The results obtained from the approach showed good improvement on clustering accuracy and algorithm performance over the other techniques where centroids convergence represents a major clustering criteria. C# and Microsoft Excel were the software used in the approach.</p>

Pendekatan Initial Centroid Search Untuk Meningkatkan Efisiensi Iterasi Klustering K-Means

Pengelompokan K-Means bertujuan untuk mengumpulkan satu set titik pusat cluster yang optimal melalui iterasi yang berurutan. Fakta bahwa semakin optimal posisi dari titik pusat awal maka semakin sedikit jumlah iterasi dari algoritma pengelompokkan K-Means untuk konvergen. Oleh karena itu, Salah satu cara untuk menemukan set initial centroid adalah melalui metode iteratif guna mencari sejumlah initial centroid yang lebih baik untuk proses pengelompokan K-Means. Langkah awal yang kami lakukan adalah mengambil sampel data dari set data dan menjalankan algoritma K-Means sebagai proses awal untuk inisialisasi centroid cluster. Kemudian kami mengulang proses iterasi dengan sejumlah initial centroid yang telah diinisialisasikan sebelumnya dan mengukur hasil pengelompokkan melalui sum-of-square-error guna menentukan kebaikan keanggotaan cluster. Centroid akhir yang memberikan jarak terendah yang akan kami teruskan ke proses pengelompokan K-means secara lengkap. Harapan kami adalah pendekatan ini akan mengarah pada set initial centroid yang lebih baik sebagai proses pengelompokan K-Means sehingga mampu meningkatkan kinerja Algoritma K-Means karena hasil konvergensi Algoritma K-Means akan berbanding lurus dengan pemilihan initial centroid.

Analisis Metode Clustering Pada Wilayah Berdasarkan Jenis Pencemaran Lingkungan Hidup

This study uses a grouping model in determining areas based on the type of environmental pollution. This study is a special reference from the government in improving environmental sustainability. The data from this study was taken from the website of the government statistical data provider, BPS (Statistics Indonesia) www.bps.go.id. This research uses the K-Mens method and releases it with RapidMiner software to create 2 clusters, high and low level clusters and see what the contents of the cluster are. From the research results obtained by high cluster centroid data that is ((1527), (810.4), (5865), (6655.3), (323), (315.1)) low cluster namely ((139.25) , (122.5), (508,833), (919,222), (64,417), (94,444)). With this analysis, it is expected to be able to load and information for the government to pay more attention to regions whose income is still below average.

K-Means Clustering-Based Electrical Equipment Identification for Smart Building Application

With the development and popular application of Building Internet of Things (BIoT) systems, numerous types of equipment are connected, and a large volume of equipment data is collected. For convenient equipment management, the equipment should be identified and labeled. Traditionally, this process is performed manually, which not only is time consuming but also causes unavoidable omissions. In this paper, we propose a k-means clustering-based electrical equipment identification toward smart building application that can automatically identify the unknown equipment connected to BIoT systems. First, load characteristics are analyzed and electrical features for equipment identification are extracted from the collected data. Second, k-means clustering is used twice to construct the identification model. Preliminary clustering adopts traditional k-means algorithm to the total harmonic current distortion data and separates equipment data into two to three clusters on the basis of their electrical characteristics. Later clustering uses an improved k-means algorithm, which weighs Euclidean distance and uses the elbow method to determine the number of clusters and analyze the results of preliminary clustering. Then, the equipment identification model is constructed by selecting the cluster centroid vector and distance threshold. Finally, identification results are obtained online on the basis of the model outputs by using the newly collected data. Successful applications to BIoT system verify the validity of the proposed identification method.