Initial Centroids Research Articles

A Machine Learning Approach to Extract Rock Mass Discontinuity Orientation and Spacing, from Laser Scanner Point Clouds

This study wants to give a contribution to the semi-automatic evaluation of rock mass discontinuities, orientation and spacing, as important parameters used in Engineering. In complex and inaccessible study areas, a traditional geological survey is hard to conduct, therefore, remote sensing techniques have proven to be a very useful tool for discontinuity analysis. However, critical expert judgment is necessary to make reliable analyses. For this reason, the open-source Python tool named DCS (Discontinuities Classification and Spacing) was developed to manage point cloud data. The tool is written in Python and is based on semi-supervised clustering. By this approach the users can: (a) estimate the number of discontinuity sets (here referred to as “clusters”) using the Error Sum of Squares (SSE) method and the K-means algorithm; (b) evaluate step by step the quality of the classification visualizing the stereonet and the scatterplot of dip vs. dip direction from the clustering; (c) supervise the clustering procedure through a manual initialization of centroids; (d) calculate the normal spacing. In contrast to other algorithms available in the literature, the DCS method does not require complex parameters as inputs for the classification and permits the users to supervise the procedure at each step. The DCS approach was tested on the steep coastal cliff of Ancona town (Italy), called the Cardeto–Passetto cliff, which is characterized by a complex fracturing and is largely affected by rockfall phenomena. The results of discontinuity orientation were validated with the field survey and compared with the ones of the FACETS plug-in of CloudCompare. In addition, the algorithm was tested and validated on regular surfaces of an anthropic wall located at the bottom of the cliff. Eventually, a kinematic analysis of rock slope stability was performed, discussing the advantages and limitations of the methods considered and making fundamental considerations on their use.

Investigation on the bouncing and coalescence behaviors of bubble pairs based on an improved APR-SPH method

Bubbles rising in fluids are widespread in many engineering fields. In this paper, the bouncing and coalescing of two bubbles rising side by side are studied based on a multi-phase smoothed particle hydrodynamics (SPH) model. An improved adaptive particle refinement (APR) scheme is applied in the SPH model to make the numerical simulation sufficiently accurate and efficient. The effect of the surface-tension model is firstly verified, then various testing cases including the rising of a single bubble and two horizontally positioned bubbles are carried out and validate the numerical model well. Furthermore, the interaction between two three-dimensional bubbles rising side by side is studied considering the effects of viscous forces (with the Reynolds number Re), surface tensions (with the Eotvos number Eo), and the initial centroid distance of the two bubbles (Γ). The motion and the deformation of the bubbles are displayed and the mechanism of the bubbles' interaction is discussed. The results reveal that for the lower Eotvos number cases indicating a larger surface tension effect, the surface tension force increases dramatically when the two bubbles collide with each other and coalesce.

Min–max kurtosis mean distance based k-means initial centroid initialization method for big genomic data clustering

Min–max kurtosis mean distance based k-means initial centroid initialization method for big genomic data clustering

K-Means Performance Optimization Using Rank Order Centroid (ROC) And Braycurtis Distance

K-Means is a clustering algorithm that groups data based on similarities between data. Some of the problems that arise from this algorithm are when determining the center point of the cluster randomly. This will certainly affect the final result of a clustering process. To anticipate the poor accuracy value, a process is needed to determine the initial centroid in the initialization process. The second problem is when calculating the Euclidean distance on the distance between data. However, this method only gives the same impact on each data attribute. From some of these problems, this study proposes the Rank Order Centroid (ROC) method for initializing the cluster center point and using the Braycurtis distance method to calculate the distance between data. With the experiment K=2 to K=10, the results obtained in this study are the proposed method obtains an iteration reduction of 6.6% on the Student Performance Exams dataset and 19.3% on the Body Fat Prediction dataset. However, there was an increase in iterations on the Heart Failure dataset by 24.2%. In testing the cluster results using the Silhouette Coefficient, this method shows an increase in the evaluation value of 5.9% in the Student Performance Exams dataset. However, the evaluation value decreased by 8.3% in the Body Fat Prediction dataset and 3.3% in the Heart Failure dataset.

From machine learning to deep learning: experimental comparison of machine learning and deep learning for skin cancer image segmentation

Skin lesion analysis is a tedious and challenging task, thus, in this research the suitability of employing machine learning or deep learning approaches for automatic lesion segmentation on dermoscopic skin cancer images is determined. The segmented region can assist clinical experts in understanding the complex lesion structure and internal pattern to find the correct skin cancer type for its early diagnosis and prevention. In this study, I present two methodologies for performing lesion segmentation: machine learning-based optimized K-means with Firefly Algorithm (FA) and Convolutional Neural Network (CNN). In the first model, the FA is hybridized with K-means clustering based on the novel average intensity fitness function to optimize the segmentation map. It is observed in the experimental results that the K-means algorithm may lead to poor results due to the wrong selection of initial centroid value, thus FA is hybridized into it to improve the performance. The second model is an enhanced encoder-decoder-based CNN framework implemented in an end-to-end fashion. These two models are compared to understand whether machine learning or deep learning is suitable to perform medical image segmentation based on a few performance metrics such as accuracy, Intersection over Union (IoU), and DICE index. These methods are evaluated and compared on two benchmark datasets provided by the International Skin Imaging Collaboration (ISIC) named ISIC 2016 [1] and ISIC 2017 [2]. Experimental results showed that the CNN model outperformed the machine learning model with an accuracy difference of 7.98% on ISIC 2016, and 7.32% on ISIC 2017. I concluded from the experimental findings that the deep learning model is more accurate and efficient in segmenting the lesion area as compared to the machine learning model. Thus, findings from this experimental work will be considered for the design of an automatic classification system by incorporating a deep learning-based segmentation approach as a pre-processing step.

A hybrid k-prototypes clustering approach with improved sine-cosine algorithm for mixed-data classification

When dealing a classification problem with mixed data, most of conventional supervised learning algorithms cannot perform well due to their numerical characteristics. However, some clustering algorithms, such as k-prototypes algorithm, show their potential in clustering mixed data. Therefore, the current study intends to develop a novel clustering-based classification algorithm for mixed data to have both merits of classification and clustering. The proposed algorithm employs a sine-cosine algorithm (SCA) to find attribute weights and initial centroids for a k-prototypes algorithm. The objective function of the algorithm is formulated as a sum-up purity. To have better performance for SCA, a mutation strategy, containing Gaussian mutation, Cauchy mutation, Levy mutation, and single-point mutation, is embedded into the original SCA. The proposed algorithm is compared with some metaheuristic-based classification algorithms and existing classification algorithms. Based on the 10 data sets from UCI, the experimental results indicated that the proposed algorithm can achieve superior classification performance in terms of accuracy and Cohen's Kappa. In addition, mutation mechanism can make SCA have better performance.

An application analysis method based on Clustering

AbstractDepending upon the load intensity and time variation information, we established the indicator system to analyze the of commercial halls, so that the further optimization actions can be taken in practical application scenarios. The representative clustering algorithms including the k-means and AC algorithm are chosen for the comparative analysis, and on this basis, an enhanced k-means algorithm is put forward in this study because the traditional method is sensitive to the initial clustering centroids and lacks the stability. A series of experiments are performed in this work, and the experimental findings reveal the validity of the proposed method. It provides a new idea for the efficiency analysis of commercial halls.

SOS-SDP: An Exact Solver for Minimum Sum-of-Squares Clustering

The minimum sum-of-squares clustering problem (MSSC) consists of partitioning n observations into k clusters in order to minimize the sum of squared distances from the points to the centroid of their cluster. In this paper, we propose an exact algorithm for the MSSC problem based on the branch-and-bound technique. The lower bound is computed by using a cutting-plane procedure in which valid inequalities are iteratively added to the Peng–Wei semidefinite programming (SDP) relaxation. The upper bound is computed with the constrained version of k-means in which the initial centroids are extracted from the solution of the SDP relaxation. In the branch-and-bound procedure, we incorporate instance-level must-link and cannot-link constraints to express knowledge about which data points should or should not be grouped together. We manage to reduce the size of the problem at each level, preserving the structure of the SDP problem itself. To the best of our knowledge, the obtained results show that the approach allows us to successfully solve, for the first time, real-world instances up to 4,000 data points.

CLUSTERING OF CLOTHING SALES DATA AT TOP STORE USING K-MEANS METHOD

In the era of globalization, the development of technological sophistication is growing rapidly which is an aspect that can be utilized to achieve convenience, especially in the flow of information. This technological sophistication by all accounts is increasingly spreading with the use of computers which are currently very popular in various areas of life. For example in the fields of education, entertainment, health, especially in the business sector. Top Store is a store that is engaged in selling clothes, however, of the various types of clothes that are sold, of course, not all of them are selling very well, and some are not selling well. Sales data, purchase of goods and unexpected expenses at Top Shop is not structured, so the data are only serves as an archive for the store and not be utilized for the development strategy of marketing. Therefore, it is necessary to apply Clustering of Clothing Sales Data in Top Stores with the K- Means Method . The K-means method can be applied to Top Stores to determine which clothes are selling very well, selling well and not selling well. The application of the K-Means method in Top Stores, namely by grouping clothing stock data. Then choose 3 clusters randomly as the initial centroid. After the data in each cluster does not change, it can be seen that the final result is that there are 21 best-selling articles, 17 articles that are selling well and 12 articles that are not selling well. Then applying the K-means method to Rapidminer is done by entering product stock data, namely initial stock, sold stock and final stock which will become a database on Ms. Excel, the data is then connected to the Rapidminer Tools , and will be processed and formed K-means.

Data Mining Technique in Detecting and Predicting Cyber In Marketplace Sector

Marketplace is one business solution that can be profitable, because it is not bound by time and place. However, marketplace can be misused by irresponsible parties, and can harm others. Then a pattern is needed to predict cybercrime in order to prevent it. To get a pattern, we can use data mining. This paper presents a general idea about the model of Data Mining techniques and diverse cybercrimes in market place applications. This paper implements data mining techniques like K-Means, Influenced Association Classifier and J48 Prediction tree for investigating the cybercrime data sets. K-means selects the initial centroids so that the classifier can mine the record and also formulate predictions of cybercrimes with J48 algorithm. The knowledge of K-Means, Influenced Association Classifier and also J48 Prediction tree tends certainly to afford a enhanced, incorporated, and precise result over the cybercrime prediction in the marketplace sectors and prevent the cybercrime.

Thermalization of nuclear matter in heavy-ion collisions at Fermi energies

We analyze the time evolution of the kinetic properties of nuclear matter produced in heavy-ion collisions at Fermi energies. The collision system is simulated using Constrained Molecular Dynamics (CoMD) transport calculations whose output is the isospin, position, and momentum of the nucleons. Focusing on central 35A⋅MeV 40Ca+40Ca collisions we utilize this information to extract localized momentum distributions in volume elements of 8fm3 and time steps of 5fm/c. We then parameterize the single-particle momentum distributions with thermally motivated fit functions in the local rest frame of each cell. While the transverse-momentum distributions are well reproduced by thermal ones, the longitudinal ones carry a marked imprint of the initial nuclear motion which we capture by introducing a centroid motion into our fit functions. In particular, we find that Fermi distributions yield significantly better fits than Boltzmann ones, a consequence of the Pauli blocking implemented in CoMD. From the fits we extract the time dependence of the thermodynamic and collective properties of the excited nuclear medium. We find that the transverse temperature gradually rises to about 6MeV, which is accompanied by a dissipation of the initial centroid motion of the incoming nuclei which vanishes at about 100fm/c after initial impact. We are therefore able to track the transition of beam energy into random kinetic energy for nucleons, suggesting a three-dimensional equilibration of energy in the late stages of the collision.

Personalized Music Recommendation Algorithm Based on Spark Platform.

Aiming at the shortcomings of traditional recommendation algorithms in dealing with large-scale music data, such as low accuracy and poor real-time performance, a personalized recommendation algorithm based on the Spark platform is proposed. The algorithm is based on the Spark platform. The K-means clustering model between users and music is constructed using an AFSA (artificial fish swarm algorithm) to optimize the initial centroids of K-means to improve the clustering effect. Based on the scoring relationship between users and users and users and music attributes, the collaborative filtering algorithm is applied to calculate the correlation between users to achieve accurate recommendations. Finally, the performance of the designed recommendation model is validated by deploying the recommendation model on the Spark platform using the Yahoo Music dataset and online music platform dataset. The experimental results show that the use of improved AFSA can complete the optimization of K-means clustering centroids with good clustering results; combined with the distributed fast computing capability of Spark platform with multiple nodes, the recommendation accuracy has better performance than traditional recommendation algorithms; especially when dealing with large-scale music data, the recommendation accuracy and real-time performance are higher, which meet the current demand of personalized music recommendation.

Performance of K-means algorithm based an ensemble learning

K-means is an iterative algorithm used with clustering task. It has more characteristics such as simplicity. In the same time, it suffers from some of drawbacks, sensitivity to initial centroid values that may produce bad results, they are based on the initial centroids of clusters that would be selected randomly. More suggestions have been given in order to overcome this problem. Ensemble learning is a method used in clustering; multiple runs are executed that produce different results for the same data set. Then the final results are driven. According to this hypothesis, more ensemble learning techniques have been suggested to deal with the clustering problem. One of these techniques is "Three ways method". However, in this paper, three ways method as an ensemble technique would be suggested to be merged with k-mean algorithm in order to improve its performance and reduce the impact of initial centroids on results. Then it was compared with traditional k-means results through practical work that was executed using popular data set. The evaluation of the hypothesis was done through computing related metrics.

Gradient-based elephant herding optimization for cluster analysis.

Clustering analysis is essential for obtaining valuable information from a predetermined dataset. However, traditional clustering methods suffer from falling into local optima and an overdependence on the quality of the initial solution. Given these defects, a novel clustering method called gradient-based elephant herding optimization for cluster analysis (GBEHO) is proposed. A well-defined set of heuristics is introduced to select the initial centroids instead of selecting random initial points. Specifically, the elephant optimization algorithm (EHO) is combined with the gradient-based algorithm GBO for assigning initial cluster centers across the search space. Second, to overcome the imbalance between the original EHO exploration and exploitation, the initialized population is improved by introducing Gaussian chaos mapping. In addition, two operators, i.e., random wandering and variation operators, are set to adjust the location update strategy of the agents. Nine datasets from synthetic and real-world datasets are adopted to evaluate the effectiveness of the proposed algorithm and the other metaheuristic algorithms. The results show that the proposed algorithm ranks first among the 10 algorithms. It is also extensively compared with state-of-the-art techniques, and four evaluation criteria of accuracy rate, specificity, detection rate, and F-measure are used. The obtained results clearly indicate the excellent performance of GBEHO, while the stability is also more prominent.

Maxmin distance sort heuristic-based initial centroid method of partitional clustering for big data mining

Maxmin distance sort heuristic-based initial centroid method of partitional clustering for big data mining

RECOMMENDATION ALGORITHM USING DATA CLUSTERING

Recommender systems play a vital role in the marketing of various goods and services. Despite the intensive growth of the theory of recommendation algorithms and a large number of their implementations, many issues remain unresolved; in particular, scalability, quality of recommendations in conditions of sparse data, and cold start. A modified collaborative filtering algorithm based on data clustering with the dynamic determination of the number of clusters and initial centroids has been developed. Data clustering is performed using the k-means method and is applied to group similar users aimed at increase of the quality of the recommendation results. The number of clusters is calculated dynamically using the silhouette method, the determination of the initial centroids is not random, but relies on the number of clusters. This approach increases the performance of the recommender system and increases the accuracy of recommendations since the search for recommendations will be carried out within one cluster where all elements are already similar. Recommendation algorithms are software-implemented for the movie recommendation system. The software implementation of various methods that allow the user to receive a recommendation for a movie meeting their preferences is carried out: a modified algorithm, memory and neighborhood-based collaborative filtering methods. The results obtained for input data of 100, 500 and 2500 users under typical conditions, data sparsity and cold start were analyzed. The modified algorithm shows the best results – from 35 to 80 percent of recommendations that meet the user's expectations. The drop in the quality of recommendations for the modified algorithm is less than 10 per cent when the number of users increases from 100 to 2500, which indicates a good level of scalability of the developed solution. In the case of sparse data (40 percent of information is missing), the quality of recommendations is 60 percent. A low quality (35 percent) of recommendations was obtained in the case of a cold start – this case needs further investigation. Constructed algorithms can be used in rating recommender systems with the ability to calculate averaged scores for certain attributes. The modified recommendation algorithm is not tied to this subject area and can be integrated into other software systems.

Max stable set problem to found the initial centroids in clustering problem

In this paper, we propose a new approach to solve the document-clustering using the K-Means algorithm. The latter is sensitive to the random selection of the k cluster centroids in the initialization phase. To evaluate the quality of K-Means clustering we propose to model the text document clustering problem as the max stable set problem (MSSP) and use continuous Hopfield network to solve the MSSP problem to have initial centroids. The idea is inspired by the fact that MSSP and clustering share the same principle, MSSP consists to find the largest set of nodes completely disconnected in a graph, and in clustering, all objects are divided into disjoint clusters. Simulation results demonstrate that the proposed K-Means improved by MSSP (KM_MSSP) is efficient of large data sets, is much optimized in terms of time, and provides better quality of clustering than other methods.

An Extensive Empirical Comparison of k-means Initialization Algorithms

The k-means clustering algorithm, whilst widely popular, is not without its drawbacks. In this paper, we focus on the sensitivity of k-means to its initial set of centroids. Since the cluster recovery performance of k-means can be improved by better initialisation, numerous algorithms have been proposed aiming at producing good initial centroids. However, it is still unclear which algorithm should be used in any particular clustering scenario. With this in mind, we compare 17 such algorithms on 6,000 synthetic and 28 real-world data sets. The synthetic data sets were produced under different configurations, allowing us to show which algorithm excels in each scenario. Hence, the results of our experiments can be particularly useful for those considering k-means for a non-trivial clustering scenario.

A Supervoxel Segmentation Method With Adaptive Centroid Initialization for Point Clouds

Supervoxels find applications as a pre-processing step in many image processing problems due to their ability to present a regional representation of points by correlating them into a set of clusters. Besides reducing the overall computational time for subsequent algorithms, the desirable properties in supervoxels are adherence to object boundaries and compactness. Existing supervoxel segmentation methods define the size of a supervoxel based on a user inputted resolution value. A fixed resolution results in poor performance in point clouds with non-uniform density. Whereas, other methods, in their quest for better boundary adherence, produce supervoxels with irregular shapes and elongated boundaries. In this article, we propose a new supervoxel segmentation method, based on k-means algorithm, with dynamic cluster seed initialization to ensure uniform distribution of cluster seeds in point clouds with variable densities. We also propose a new cluster seed initialization strategy, based on histogram binning of surface normals, for better boundary adherence. Our algorithm is parameter-free and gives equal importance to the color, spatial location and orientation of the points resulting in compact supervoxels with tight boundaries. We test the efficacy of our algorithm on a publicly available point cloud dataset consisting of 1449 pairs of indoor RGB-D images, i.e., color (RGB) images coupled with depth information (D) mapped per pixel. Results are compared against three state-of-the-art algorithms based on four quality metrics. Results show that our method provides significant improvement over other methods in the undersegmentation error and compactness metrics and, performs equally well in the boundary recall and contour density metrics.

Image Segmentation Using PSO-Enhanced K-Means Clustering and Region Growing Algorithms

Image segmentation is a basic image processing technique that is primarily used for finding segments that form the entire image. These segments can be then utilized in discriminative feature extraction, image retrieval, and pattern recognition. Clustering and region growing techniques are the commonly used image segmentation methods. K-Means is a heavily used clustering technique due to its simplicity and low computational cost. However, K-Means results depend on the initial centres’ values which are selected randomly, which leads to inconsistency in the image segmentation results. In addition, the quality of the isolated regions depends on the homogeneity of the resulted segments. In this paper, an improved K-Means clustering algorithm is proposed for image segmentation. The presented method uses Particle Swarm Intelligence (PSO) for determining the initial centres based on Li’s method. These initial centroids are then fed to the K-Means algorithm to assign each pixel into the appropriate cluster. The segmented image is then given to a region growing algorithm for regions isolation and edge map generation. The experimental results show that the proposed method gives high quality segments in a short processing time.