Well-known Clustering Techniques Research Articles

Insurance Analytics with Clustering Techniques

The K-means algorithm and its variants are well-known clustering techniques. In actuarial applications, these partitioning methods can identify clusters of policies with similar attributes. The resulting partitions provide an actuarial framework for creating maps of dominant risks and unsupervised pricing grids. This research article aims to adapt well-established clustering methods to complex insurance datasets containing both categorical and numerical variables. To achieve this, we propose a novel approach based on Burt distance. We begin by reviewing the K-means algorithm to establish the foundation for our Burt distance-based framework. Next, we extend the scope of application of the mini-batch and fuzzy K-means variants to heterogeneous insurance data. Additionally, we adapt spectral clustering, a technique based on graph theory that accommodates non-convex cluster shapes. To mitigate the computational complexity associated with spectral clustering’s O(n3) runtime, we introduce a data reduction method for large-scale datasets using our Burt distance-based approach.

Density peaks clustering based on density voting and neighborhood diffusion

Density Peaks Clustering (DPC) is a well-known clustering technique in the data mining field with fewer parameters as well as no iteration. However, when dealing with datasets containing multiple peaks, DPC may subjectively choose the wrong cluster centers through the decision graph. Additionally, DPC requires a considerable amount of time to estimate density and relative distance. Moreover, DPC is sensitive to the value of cut-off distance. To overcome these issues, a density peaks clustering algorithm based on density voting and neighborhood diffusion (DPC-DVND) is proposed. Firstly, the proposed algorithm utilizes the k nearest neighbors and KD-tree to enhance the efficiency of computing local density and relative distance. Secondly, this study selects the potential cluster centers by density voting and applies the number of votes instead of density to calculate the feasibility of each potential center becoming a cluster center, so that the centers of low-density clusters can be better distinguished. Finally, two neighborhood density diffusion rules are designed to propagate labels and form the core structure of clusters. Experiments on synthetic, real, and image datasets are performed to compare different methods. Results show that DPC-DVND outperforms other state-of-the-art algorithms in terms of effectiveness and efficiency.

Robust clustering algorithm: The use of soft trimming approach

The presence of noise or outliers in data sets may heavily affect the performance of clustering algorithms and lead to unsatisfactory results. The majority of conventional clustering algorithms are sensitive to noise and outliers. Robust clustering algorithms often overcome difficulties associated with noise and outliers and find true cluster structures. We introduce a soft trimming approach for the hard clustering problem where its objective is modeled as a sum of the cluster function and a function represented as a composition of the algebraic and distance functions. We utilize the composite function to estimate the degree of the significance of each data point in clustering. A robust clustering algorithm based on the new model and a procedure for generating starting cluster centers is developed. We demonstrate the performance of the proposed algorithm using some synthetic and real-world data sets containing noise and outliers. We also compare its performance with that of some well-known clustering techniques. Results show that the new algorithm is robust to noise and outliers and finds true cluster structures.

Face Quality Detection in a Video Frame

Abstract: Face detection technology is often used for surveillance of detecting and tracking of people in real time. The applications using these algorithms deal with low quality video feeds having less Pixels Per Inch (ppi) and/or low frame rate. The algorithms perform well with such video feeds, but their performance deteriorates towards high quality, high data-per-frame videos. This project focuses on developing such an algorithm that gives faster results on high quality videos, at par with the algorithms working on low quality videos. The proposed algorithm uses MTCNN as base algorithm, and speeds it up for highdefinition videos. This project also presents a novel solution to the problem of occlusion and detecting faces in videos. This survey provides an overview of the face detection from video literature, which predominantly focuses on visible wavelength face video as input. For the high-quality videos, we will Face-MTCNN and KLT, for low quality videos we will use MTCNN and KLT. Open issues and challenges are pointed out, i.e., highlighting the importance of comparability for algorithm evaluations and the challenge for future work to create Deep Learning (DL) approaches that are interpretable in addition to Track the faces. The suggested methodology is contrasted with conventional facial feature extraction for every frame and with well-known clustering techniques for a collection of videos

Tempo-categorization of road accident hotspots to enhance the problem diagnosis process and detect hidden hazardous locations

Identifying roads’ hazardous locations and solving their problems are key measures in traffic safety management. However, since the traditional hotspot identification (HSID) rests on the yearly-aggregated crashes, two problems appear: locations that become unsafe at specific short periods may remain unidentified as they may not show noticeable crash counts, and results of the problem diagnosis analysis on hotspots’ crashes potentially contain a great amount of uncertainty. Even though researchers have recently added the dimension of time and analyzed accidents spatio-temporally to obtain more insights, the mentioned problems have not been addressed fully. Hence, this paper first suggests a new linear DBSCAN-based HSID method and demonstrates its acceptable performance by comparison with KDE+, the well-known clustering technique; Second, employing the proposed technique, the paper presents an algorithm for the spatial analysis of accidents through diverse time dimensions, which categorizes the risky locations based on their periodic reappearance. The tempo-categorization purpose is to enhance diagnosing causative risks by understanding their arising periods. The algorithm is tested using Allegheny highways crash data from 2014 to 2019. Results illustrate the contribution of the suggested method to problem diagnosis and detecting hidden unsafe points.

Data clustering using leaders and followers optimization and differential evolution

Data clustering is an important research topic in data mining. Although cluster analysis based on optimization algorithms has attracted great attention, optimization-based techniques face difficulties due to the non-linear objective function and complicated search space. Leaders and followers optimization (LaF) introduced in the 2015 IEEE Congress on Evolutionary Computation, and differential evolution algorithm (DE) are two efficient evolutionary computation methods, and they own some special advantages. The key power of LaF is the exploration in multi-modal search spaces, but it has a poor performance in the exploitation. On the other hand, DE based on the DE/best/1 mutation operator significantly promotes the exploitation process. In this study, the strong properties of LaF and DE are combined to balance exploration and exploitation in the search space to discover the cluster centroids. Besides, the proposed clustering method, i.e., LaF-DE, does not need parameter settings, unlike the existing optimization-based partitional clustering methods. Hence, this study proposes a straightforward, parameter-free, and efficient novel hybrid algorithm for the optimization-based partitional data clustering problem. Many experiments on the functions from CEC2017 test suite show that LaF-DE has better optimization performance and higher stability than the state-of-the-art metaheuristic algorithms. LaF-DE has been compared with well-known clustering techniques on the UCI and Shape datasets. The experimental results and statistical tests indicate that LaF-DE outperforms the well-known partitional clustering methods on 8 of 12 datasets in terms of internal performance metrics. Besides, LaF-DE performs better than density peaks clustering on 9 of 12 datasets in terms of external performance metrics.

A fuzzy clustering technique for enhancing the convergence performance by using improved Fuzzy c-means and Particle Swarm Optimization algorithms

Fuzzy clustering is a well-established technique among the well-known clustering techniques in several real-world applications due to easy implementation and produces satisfactory clustering result. However, it has some deficiency such as sensitive to outliers, result dependency on choosing initial centroid, etc. To eradicate the shortcoming of FCM algorithm, this article introduces a robust clustering technique, particle swarm optimization improved fuzzy c-means is developed by the hybridization of particle swarm optimization and improved fuzzy c-means techniques, to deal with noisy data and initialization problem. In this article, a fuzzy clustering technique is developed to increase the convergence performance of clustering techniques. Fuzzy c-means is improved by developing a new metric to tolerate the noisy environment. Particle swarm optimization has an inbuilt guidance strategy which leads the solution in particle swarm optimization to obtain useful information from the better solution and thereby helping them improve their own solution. To handle the initialization problem of fuzzy c-means, particle swarm optimization technique is used. PSO effectively enhance the performance of improved FCM to increase the effectiveness of clustering. The effectiveness of the proposed clustering technique over existing techniques in literature has been illustrated by adopting eight real worlds and three artificial data sets. The results show that the proposed algorithm generates encouraging results as compared to the established clustering technique in literature.

Clustering Using Cyclic Spaces of Reversible Cellular Automata

This paper introduces a cycle-based clustering technique using the cyclic spaces of reversible cellular automata (CAs). Traditionally, a cluster consists of close objects, which in the case of CAs necessarily means that the objects belong to the same cycle; that is, they are reachable from each other. Each of the cyclic spaces of a cellular automaton (CA) forms a unique cluster. This paper identifies CA properties based on “reachability” that make the clustering effective. To do that, we first figure out which CA rules contribute to maintaining the minimum intracluster distance. Our CA is then designed with such rules to ensure that a limited number of cycles exist in the configuration space. An iterative strategy is also introduced that can generate a desired number of clusters by merging objects of closely reachable clusters from a previous level in the present level using a unique auxiliary CA. Finally, the performance of our algorithm is measured using some standard benchmark validation indices and compared with existing well-known clustering techniques. It is found that our algorithm is at least on a par with the best algorithms existing today on the metric of these standard validation indices.

Traffic-Aware Clustering Scheme for MANET Using Modified Elephant Herding Optimization Algorithm

Clustering is the prevalent routing method in the large-scale Mobile Ad Hoc Network (MANET). The Cluster-Heads (CHs) play an important role in routing as it is transient through all communications of its associated nodes. To ensure fairness in the use of energy in all clusters, each CH has to deal with same amount of traffic. The previous clustering methods focused mainly on the distribution of equal member nodes in each cluster. They failed to consider every cluster's traffic generated. This paper introduces a novel technique for MANET clustering with Modified Elephant Herding Optimization based on the traffic generated within each cluster. This Traffic-Aware Clustering with Modified Elephants Herding Optimization (TAC-MEHO) produces optimized clusters for stable communication and is experimentally tested with well-known clustering techniques. Assessment metrics such as number of Cluster-Heads (CHs), lifetime of the network, and re-clustering rates are measured using various parameter values such as network size, network traffic and transmission distance. The results show that proposed TAC-MEHO improves the re-clustering rate by 91% and 58% when compared with Weighted Clustering Algorithm (WCA) and WCA-GA respectively. Further, it improves the network lifetime by 89% and 88 % over WCA and WCA-GA respectively.

Enhancing the Role of Large-Scale Recommendation Systems in the IoT Context

The Internet of Things (IoT) connects heterogeneous physical devices with the ability to collect data using sensors and actuators. These data can infer useful information for decision-makers in many applications systems monitoring, healthcare, transportations, data storage, smart homes, and many others. In the Era of the Internet of Things (IoT), recommender systems can support scenarios such as recommending apps, IoT workflows, services, sensor equipment, hotels, and drugs to users and customers. Current state-of-art recommendation systems, including collaborative filtering methods, suffer from scalability and sparsity problems. This article proposes a clustering-based recommendation system that adopts the vector space model from information retrieval to obtain highly accurate recommendations. The proposed algorithm uses four well-known clustering techniques as k-means (KM), fuzzy c-mean (FCM), Single-Linkage (SLINK), and Self-Organizing-Maps (SOM). Various experiments and benchmarking on seven IoT rating datasets from different fields are conducted to assess the performance of the proposed recommender system. The experimental results using both error and prediction metrics indicate that the proposed algorithm outperforms the traditional collaborative filtering approach. Besides, adopting the self-organizing strategy obtains recommendations of significant accuracy as compared to the partitional learning approaches.

K-MILP: A novel clustering approach to select typical and extreme days for multi-energy systems design optimization

When optimizing the design of multi-energy systems, the operation strategy and the part-load behavior of the units must be considered in the optimization model, which therefore must be formulated as a two-stage problem. In order to guarantee computational tractability, the operation problem is solved for a limited set of typical and extreme periods. The selection of these periods is an important aspect of the design methodology, as the selection and sizing of the units is carried out on the basis of their optimal operation in the selected periods. This work proposes a novel Mixed Integer Linear Program clustering model, named k-MILP, devised to find at the same time the most representative days of the year and the extreme days. k-MILP allows controlling the features of the selected typical and extreme days and setting a maximum deviation tolerance on the integral of the load duration curves. The novel approach is tested on the design of two different multi-energy systems (a multiple-site university Campus and a single building) and compared with the two well-known clustering techniques k-means and k-medoids. Results show that k-MILP leads to a better representation of both typical and extreme operating conditions guiding towards more efficient and reliable designs.

Novel Adaptive Clustering Algorithms Based on a Probabilistic Similarity Measure Over Atanassov Intuitionistic Fuzzy Set

This paper presents a novel probabilistic similarity measure (PSM) for Atanassov intuitionistic fuzzy sets. It then exploits PSM to propose an adaptive probabilistic similarity degree and develops the novel probabilistic $\lambda$ -cutting algorithm for clustering. Further, the probabilistic distance measure (obtained from the PSM) is used to develop a new clustering technique, which we have named “probabilistic intuitionistic fuzzy c-mean (PIFCM) algorithm”. Simulation experiments have been conducted over a variety of datasets including UCI machine learning datasets and real-world car dataset. The results obtained have been thoroughly compared with other well-known clustering techniques such as fuzzy c-mean (FCM), intuitionistic fuzzy c-mean, association coefficient method, and $\lambda$ -cutting method. Based upon the experimental results, it can be concluded that our probabilistic $\lambda$ -cutting algorithm and PIFCM algorithm outperform their existing counterparts.

Gene Expression Data Clustering Using Variance-based Harmony Search Algorithm

ABSTRACTThe advancement in microarray technology has led to a flare-up of gene expression data, thus giving rise to need of efficient technique to analyze these massive data-sets. Among the existing techniques, clustering is a well-known technique used for analysis of microarray data. In this paper, a novel metaheuristic clustering approach for analysis of gene expression is proposed. A variance-based harmony search algorithm is used as an underlying metaheuristic algorithm. The performance of the proposed approach has been tested on 6 real-life data-sets and compared with 12 well-known clustering techniques. The experimental results reveal that the proposed technique outperforms the other existing techniques. The statistical and biological significance tests have also been carried out to demonstrate the superiority of the proposed technique.

Automatic clustering and feature selection using gravitational search algorithm and its application to microarray data analysis

This paper proposes a novel approach that selects the number of clusters along with relevant features automatically and simultaneously. Gravitational search algorithm is used as metaheuristic. A novel agent representation scheme is used for encoding cluster centers and number of features. The algorithm is able to find the optimal number of clusters and the relevant features corresponding to the clusters during the run time. A new concept of threshold setting is used. The variance (statistical property) of the dataset has been exploited. To make the search efficient, a novel clustering criterion is used. The proposed approach is compared with recently developed well-known clustering techniques. This approach is further applied for analysis of microarray data. The statistical and biological significance tests are performed to demonstrate the efficiency of proposed approach. The results prove the effectiveness and the accuracy of the proposed algorithm.

Optimal energy aware clustering in circular wireless sensor networks

The lifetime of a wireless sensor node refers to the duration after which the node's energy has ended. Since battery replacement in most applications of wireless sensor networks is not possible, designing an energy-efficient communication protocol in these networks is very important. Therefore, many studies have been conducted to find a solution to increase the lifetime of these networks. Clustering is a useful technique for partitioning the network to areas, called clusters and entrusting energy-waste issues (e.g. data gathering, aggregating and routing to the sink) to some specific nodes, the cluster heads. In this paper, a new method for Optimal Clustering in Circular Networks (OCCN) is proposed which aims to mitigate energy consumption and increase the lifetime of wireless sensor networks. In this method, which is proposed for a circular area surrounding a sink, one hop communication between the cluster heads and the sink is replaced by an optimal multi-hop communications. Moreover, the optimal number of clusters is computed and the energy consumption is optimized by partitioning the network into nearly the same size clusters in a distributed manner. Simulation results indicate that the proposed method achieved more than 35% improvements in terms of energy consumption in comparison to other well-known clustering techniques.

A novel hybridization strategy for krill herd algorithm applied to clustering techniques

Krill herd (KH) is a stochastic nature-inspired optimization algorithm that has been successfully used to solve numerous complex optimization problems. This paper proposed a novel hybrid of KH algorithm with harmony search (HS) algorithm, namely, H-KHA, to improve the global (diversification) search ability. The enhancement includes adding global search operator (improvise a new solution) of the HS algorithm to the KH algorithm for improving the exploration search ability by a new probability factor, namely, Distance factor, thereby moving krill individuals toward the best global solution. The effectiveness of the proposed H-KHA is tested on seven standard datasets from the UCI Machine Learning Repository that are commonly used in the domain of data clustering, also six common text datasets that are used in the domain of text document clustering. The experiments reveal that the proposed hybrid KHA with HS algorithm (H-KHA) enhanced the results in terms of accurate clusters and high convergence rate. Mostly, the performance of H-KHA is superior or at least highly competitive with the original KH algorithm, well-known clustering techniques and other comparative optimization algorithms.

A recommender system for tourism industry using cluster ensemble and prediction machine learning techniques

Recommender systems have emerged in the e-commerce domain and are developed to actively recommend the right items to online users. Traditional Collaborative Filtering (CF) recommender systems recommend the items to users based on their single-rating feedback which are used to match similar users. In multi-criteria CF recommender systems, however, multi-criteria ratings are used instead of single-rating feedback which can significantly improve the accuracy of traditional CF algorithms. These systems have been successfully implemented in Tourism domain. In this paper, we propose a new recommendation method based on multi-criteria CF to enhance the predictive accuracy of recommender systems in tourism domain using clustering, dimensionality reduction and prediction methods. We use Adaptive Neuro-Fuzzy Inference Systems (ANFIS) and Support Vector Regression (SVR) as prediction techniques, Principal Component Analysis (PCA) as a dimensionality reduction technique and Self-Organizing Map (SOM) and Expectation Maximization (EM) as two well-known clustering techniques. To improve the recommendation accuracy of proposed multi-criteria CF, a cluster ensembles approach, Hypergraph Partitioning Algorithm (HGPA), is applied on SOM and EM clustering results. We evaluate the accuracy of recommendation method on TripAdvisior dataset. Our experiments confirm that cluster ensembles can provide better predictive accuracy for the proposed recommendation method in relation to the methods which solely rely on single clustering techniques.

Enhancing point symmetry-based distance for data clustering

In this paper, at first a new point symmetry-based similarity measurement is proposed which satisfies the closure and the symmetry properties of any distance function. The different desirable properties of the new distance are elaborately explained. Thereafter a new clustering algorithm based on the search capability of genetic algorithm is developed where the newly developed point symmetry-based distance is used for cluster assignment. The allocation of points to different clusters is performed in such a way that the closure property is satisfied. The proposed GA with newly developed point symmetry distance based (GAnPS) clustering algorithm is capable of determining different symmetrical shaped clusters having any sizes or convexities. The effectiveness of the proposed GAnPS clustering technique in identifying the proper partitioning is shown for twenty-one data sets having various characteristics. Performance of GAnPS is compared with existing symmetry-based genetic clustering technique, GAPS, three popular and well-known clustering techniques, K-means, expectation maximization and average linkage algorithm. In a part of the paper, the utility of the proposed clustering technique is shown for partitioning a remote sensing satellite image. The last part of the paper deals with the development of some automatic clustering techniques using the newly proposed symmetry-based distance.

A Fuzzy set-based method to identify the car position in a road lane at intersections by smartphone GPS data

Abstract Intelligent transportation systems (ITS) work by collections of data in real time. Average speed, travel time and delay at intersections are some of the most important measures, often used for monitoring the performance of transportation systems, and useful for system management and planning. In urban transportation planning, intersections are usually considered critical points, acting as bottlenecks and clog points for urban traffic. Thus, detecting the travel time at intersections in different turning directions is an activity useful to improve the urban transport efficiency. Smartphones represent a low-cost technology, with which is possible to obtain information about traffic state. However, smartphone GPS data suffer for low precision, mainly in urban areas. In this paper, we present a fuzzy set-based method for car positioning identification within road lanes near intersections using GPS data coming from smartphones. We have introduced the fuzzy sets to take into account uncertainty embedded in GPS data when trying to identify the position of cars within the road lanes. Moreover, we introduced a Genetic Algorithm to calibrate the fuzzy parameters in order to obtain a novel supervised clustering technique. We applied the proposed method to one intersection in the urban road network of Bari (Italy). First results reveal the effectiveness of the proposed methodology when comparing the outcomes of the proposed method with two well-known clustering techniques (Fuzzy C-means, K-means).

Performance Evaluation of Line Symmetry-Based Validity Indices on Clustering Algorithms

AbstractFinding the optimal number of clusters and the appropriate partitioning of the given dataset are the two major challenges while dealing with clustering. For both of these, cluster validity indices are used. In this paper, seven widely used cluster validity indices, namely DB index, PS index, I index, XB index, FS index, K index, and SV index, have been developed based on line symmetry distance measures. These indices provide the measure of line symmetry present in the partitioning of the dataset. These are able to detect clusters of any shape or size in a given dataset, as long as they possess the property of line symmetry. The performance of these indices is evaluated on three clustering algorithms: K-means, fuzzy-C means, and modified harmony search-based clustering (MHSC). The efficacy of symmetry-based validity indices on clustering algorithms is demonstrated on artificial and real-life datasets, six each, with the number of clusters varying from 2 to $\sqrt n ,$ where n is the total number of data points existing in the dataset. The experimental results reveal that the incorporation of line symmetry-based distance improves the capabilities of these existing validity indices in finding the appropriate number of clusters. Comparisons of these indices are done with the point symmetric and original versions of these seven validity indices. The results also demonstrate that the MHSC technique performs better as compared to other well-known clustering techniques. For real-life datasets, analysis of variance statistical analysis is also performed.