Fault tolerant decentralised K-Means clustering for asynchronous large-scale networks

The K-Means algorithm for cluster analysis is one of the most influential and popular data mining methods. Its straightforward parallel formulation is well suited for distributed memory systems with reliable interconnection networks, such as massively parallel processors and clusters of workstations. However, in large-scale geographically distributed systems the straightforward parallel algorithm can be rendered useless by a single communication failure or high latency in communication paths. The lack of scalable and fault tolerant global communication and synchronisation methods in large-scale systems has hindered the adoption of the K-Means algorithm for applications in large networked systems such as wireless sensor networks, peer-to-peer systems and mobile ad hoc networks. This work proposes a fully distributed K-Means algorithm (EpidemicK-Means) which does not require global communication and is intrinsically fault tolerant. The proposed distributed K-Means algorithm provides a clustering solution which can approximate the solution of an ideal centralised algorithm over the aggregated data as closely as desired. A comparative performance analysis is carried out against the state of the art sampling methods and shows that the proposed method overcomes the limitations of the sampling-based approaches for skewed clusters distributions. The experimental analysis confirms that the proposed algorithm is very accurate and fault tolerant under unreliable network conditions (message loss and node failures) and is suitable for asynchronous networks of very large and extreme scale.

Efficient group communication for large-scale parallel clustering

Global communication requirements and load imbalance of some parallel data mining algorithms are the major obstacles to exploit the computational power of large-scale systems. This work investigates how non-uniform data distributions can be exploited to remove the global communication requirement and to reduce the communication cost in iterative parallel data mining algorithms. In particular, the analysis focuses on one of the most influential and popular data mining methods, the k-means algorithm for cluster analysis. The straightforward parallel formulation of the k-means algorithm requires a global reduction operation at each iteration step, which hinders its scalability. This work studies a different parallel formulation of the algorithm where the requirement of global communication can be relaxed while still providing the exact solution of the centralised k-means algorithm. The proposed approach exploits a non-uniform data distribution which can be either found in real world distributed applications or can be induced by means of multi-dimensional binary search trees. The approach can also be extended to accommodate an approximation error which allows a further reduction of the communication costs.

Large-scale dynamics associated with clustering of extratropical cyclones affecting Western Europe

Some recent winters in Western Europe have been characterized by the occurrence of multiple extratropical cyclones following a similar path. The occurrence of such cyclone clusters leads to large socio-economic impacts due to damaging winds, storm surges, and floods. Recent studies have statistically characterized the clustering of extratropical cyclones over the North Atlantic and Europe and hypothesized potential physical mechanisms responsible for their formation. Here we analyze 4 months characterized by multiple cyclones over Western Europe (February 1990, January 1993, December 1999, and January 2007). The evolution of the eddy driven jet stream, Rossby wave-breaking, and upstream/downstream cyclone development are investigated to infer the role of the large-scale flow and to determine if clustered cyclones are related to each other. Results suggest that optimal conditions for the occurrence of cyclone clusters are provided by a recurrent extension of an intensified eddy driven jet toward Western Europe lasting at least 1 week. Multiple Rossby wave-breaking occurrences on both the poleward and equatorward flanks of the jet contribute to the development of these anomalous large-scale conditions. The analysis of the daily weather charts reveals that upstream cyclone development (secondary cyclogenesis, where new cyclones are generated on the trailing fronts of mature cyclones) is strongly related to cyclone clustering, with multiple cyclones developing on a single jet streak. The present analysis permits a deeper understanding of the physical reasons leading to the occurrence of cyclone families over the North Atlantic, enabling a better estimation of the associated cumulative risk over Europe.

Serial clustering of extratropical cyclones

The clustering in time (seriality) of extratropical cyclones is responsible for large cumulative insured losses in western Europe, though surprisingly little scientific attention has been given to this important property. This study investigates and quantifies the seriality of extratropical cyclones in the Northern Hemisphere using a point-process approach. A possible mechanism for serial clustering is the time-varying effect of the large-scale flow on individual cyclone tracks. Another mechanism is the generation by one parent cyclone of one or more offspring through secondary cyclogenesis. A long cyclone-track database was constructed for extended October March winters from 1950 to 2003 using 6-h analyses of 850-mb relative vorticity derived from the NCEP NCAR reanalysis. A dispersion statistic based on the varianceto- mean ratio of monthly cyclone counts was used as a measure of clustering. It reveals extensive regions of statistically significant clustering in the European exit region of the North Atlantic storm track and over the central North Pacific. Monthly cyclone counts were regressed on time-varying teleconnection indices with a log-linear Poisson model. Five independent teleconnection patterns were found to be significant factors over Europe: the North Atlantic Oscillation (NAO), the east Atlantic pattern, the Scandinavian pattern, the east Atlantic western Russian pattern, and the polar Eurasian pattern. The NAO alone is not sufficient for explaining the variability of cyclone counts in the North Atlantic region and western Europe. Rate dependence on time-varying teleconnection indices accounts for the variability in monthly cyclone counts, and a cluster process did not need to be invoked.

Bayesian clustering of farm types using the mixtures model

A Bayesian method of classifying observations that are assumed to come from a number of distinct subpopulations is outlined. The method is illustrated with simulated data and applied to the classification of farms according to their level and variability of income. The resultant classification shows a greater diversity of technical charactersitics within farm types than is conventionally the case. The range of mean farm income between groups in the new classification is wider than that of the conventional method and the variability of income within groups is narrower. Results show that the highest income group in 2000 included large specialist dairy farmers and pig and poultry producers, whilst in 2001 it included large and small specialist dairy farms and large mixed dairy and arable farms. In both years the lowest income group is dominated by non-milk producing livestock farms.

Superovulation and embryo collection in nulliparous Boer goat does immunized against a recombinant ovine alpha-subunit inhibin

With the purpose of eliciting a superovulatory response, 12 adult nulliparous Boer goat does were actively immunized against a recombinant a-subunit of ovine inhibin (roIHN-alpha; two injections of 100 mg 4 weeks apart). Another 12 control Boer goat does were treated with physiological saline and acted as controls. One year later the immunized animals were boostered by the administration of another dose (100 mg) of the immunogen. Following treatment, blood samples were collected twice weekly for the periods of 16 and 12 weeks, respectively, to monitor the inhibin binding ability with the aid of a radio-tracer binding assay. Throughout the experiment, estrus detection was conducted twice daily with the aid of an aproned intact buck. From the first day after treatment to 48 h after standing estrus, ovarian activity was monitored daily by transrectal ultrasonography. On alternate estrous cycles, does were mated and 6 days later flushed transcervically to recover embryos. All goats treated with the roIHN-alpha produced antibodies reactive to the native bovine inhibin tracer-the titre increasing from 2.9 +/- 0.4 to a maximum of 21.9 +/- 2.9% binding after the second injection. The antibody titre gradually subsided over the next 16 weeks. The booster injection restored an elevated antibody titre (11.7 +/- 0.4%), which was maintained until the end of the sampling period 12 weeks later. In the control goats only trace amounts of antibody were recorded throughout the trial. In the roIHN-alpha-immunized goats the number of follicles reaching a diameter of > 4 mm was 14.6 +/- 1.2 per doe. A positive correlation was recorded between the follicle number and antibody titre (r=0.61; P < 0.01). The number of follicles ovulating per doe (6.9 +/- 0.7) followed the same tendency-however, the proportion decreased with increasing follicle numbers. A relatively weak correlation was recorded between the inhibin binding ability and number of ovulations (r=0.27; P < 0.05). In the control goats the majority (92%) of follicles exceeding 4 mm in diameter ovulated (2.5 +/- 0.1 follicles/doe). Embryo collection proved unsatisfactory (42% versus 39% recovery for immunized and control animals, respectively)-presumably because the uterine lumen of the nulliparous does was too narrow to permit effective flushing. In the group of immunized goats the occurrence of short estrous cycles (< 15 days) recorded was 34% versus only 6% in the controls. Overall, immunization of goats against roIHN-alpha led to an almost six-fold increase in number of ovarian follicles, a three-fold increase in ovulations and, despite the low recovery rate, a more than three-fold increase in ova or embryos recovered. It may be concluded that treatment of female goats with roIHN-alpha leads to an inhibin antibody response, accompanied by enhanced ovarian activity. The response was, however, accompanied by a large proportion of retained follicles and a high incidence of short estrous cycles. These problems need to be further investigated before rendering the method fit for application in embryo transfer programs in goats. (C) 2007 Elsevier B.V. All rights reserved.

Epidemic K-Means clustering

The K-Means algorithm for cluster analysis is one of the most influential and popular data mining methods. Its straightforward parallel formulation is well suited for distributed memory systems with reliable interconnection networks. However, in large-scale geographically distributed systems the straightforward parallel algorithm can be rendered useless by a single communication failure or high latency in communication paths. This work proposes a fully decentralised algorithm (Epidemic K-Means) which does not require global communication and is intrinsically fault tolerant. The proposed distributed K-Means algorithm provides a clustering solution which can approximate the solution of an ideal centralised algorithm over the aggregated data as closely as desired. A comparative performance analysis is carried out against the state of the art distributed K-Means algorithms based on sampling methods. The experimental analysis confirms that the proposed algorithm is a practical and accurate distributed K-Means implementation for networked systems of very large and extreme scale.

Non-uniform data distribution for communication-efficient parallel clustering

Global communicationrequirements andloadimbalanceof someparalleldataminingalgorithms arethe major obstacles to exploitthe computational power of large-scale systems. This work investigates how non-uniform data distributions can be exploited to remove the global communication requirement and to reduce the communication costin parallel data mining algorithms and, in particular, in the k-means algorithm for cluster analysis. In the straightforward parallel formulation of the k-means algorithm, data and computation loads are uniformly distributed over the processing nodes. This approach has excellent load balancing characteristics that may suggest it could scale up to large and extreme-scale parallel computing systems. However, at each iteration step the algorithm requires a global reduction operationwhichhinders thescalabilityoftheapproach.Thisworkstudiesadifferentparallelformulation of the algorithm where the requirement of global communication is removed, while maintaining the same deterministic nature ofthe centralised algorithm. The proposed approach exploits a non-uniform data distribution which can be either found in real-world distributed applications or can be induced by means ofmulti-dimensional binary searchtrees. The approachcanalso be extended to accommodate an approximation error which allows a further reduction ofthe communication costs. The effectiveness of the exact and approximate methods has been tested in a parallel computing system with 64 processors and in simulations with 1024 processing element

Dynamic group communication for large-scale parallel data mining

Exascale systems are the next frontier in high-performance computing and are expected to deliver a performance of the order of 10^18 operations per second using massive multicore processors. Very large- and extreme-scale parallel systems pose critical algorithmic challenges, especially related to concurrency, locality and the need to avoid global communication patterns. This work investigates a novel protocol for dynamic group communication that can be used to remove the global communication requirement and to reduce the communication cost in parallel formulations of iterative data mining algorithms. The protocol is used to provide a communication-efficient parallel formulation of the k-means algorithm for cluster analysis. The approach is based on a collective communication operation for dynamic groups of processes and exploits non-uniform data distributions. Non-uniform data distributions can be either found in real-world distributed applications or induced by means of multidimensional binary search trees. The analysis of the proposed dynamic group communication protocol has shown that it does not introduce significant communication overhead. The parallel clustering algorithm has also been extended to accommodate an approximation error, which allows a further reduction of the communication costs. The effectiveness of the exact and approximate methods has been tested in a parallel computing system with 64 processors and in simulations with 1024 processing elements.

Serial clustering of extratropical cyclones over the North Atlantic and Europe under recent and future climate conditions

Under particular large-scale atmospheric conditions, several windstorms may affect Europe within a short time period. The occurrence of such cyclone families leads to large socioeconomic impacts and cumulative losses. The serial clustering of windstorms is analyzed for the North Atlantic/western Europe. Clustering is quantified as the dispersion (ratio variance/mean) of cyclone passages over a certain area. Dispersion statistics are derived for three reanalysis data sets and a 20-run European Centre Hamburg Version 5 /Max Planck Institute Version–Ocean Model Version 1 global climate model (ECHAM5/MPI-OM1 GCM) ensemble. The dependence of the seriality on cyclone intensity is analyzed. Confirming previous studies, serial clustering is identified in reanalysis data sets primarily on both flanks and downstream regions of the North Atlantic storm track. This pattern is a robust feature in the reanalysis data sets. For the whole area, extreme cyclones cluster more than nonextreme cyclones. The ECHAM5/MPI-OM1 GCM is generally able to reproduce the spatial patterns of clustering under recent climate conditions, but some biases are identified. Under future climate conditions (A1B scenario), the GCM ensemble indicates that serial clustering may decrease over the North Atlantic storm track area and parts of western Europe. This decrease is associated with an extension of the polar jet toward Europe, which implies a tendency to a more regular occurrence of cyclones over parts of the North Atlantic Basin poleward of 50°N and western Europe. An increase of clustering of cyclones is projected south of Newfoundland. The detected shifts imply a change in the risk of occurrence of cumulative events over Europe under future climate conditions.

Hierarchical clustering based web service discovery

This paper presents a hierarchical clustering method for semantic Web service discovery. This method aims to improve the accuracy and efficiency of the traditional service discovery using vector space model. The Web service is converted into a standard vector format through the Web service description document. With the help of WordNet, a semantic analysis is conducted to reduce the dimension of the term vector and to make semantic expansion to meet the user’s service request. The process and algorithm of hierarchical clustering based semantic Web service discovery is discussed. Validation is carried out on the dataset.

On the clustering of winter storm loss events over Germany

During the last decades, several windstorm series hit Europe leading to large aggregated losses. Such storm series are examples of serial clustering of extreme cyclones, presenting a considerable risk for the insurance industry. Clustering of events and return periods of storm series for Germany are quantified based on potential losses using empirical models. Two reanalysis data sets and observations from German weather stations are considered for 30 winters. Histograms of events exceeding selected return levels (1-, 2- and 5-year) are derived. Return periods of historical storm series are estimated based on the Poisson and the negative binomial distributions. Over 4000 years of general circulation model (GCM) simulations forced with current climate conditions are analysed to provide a better assessment of historical return periods. Estimations differ between distributions, for example 40 to 65 years for the 1990 series. For such less frequent series, estimates obtained with the Poisson distribution clearly deviate from empirical data. The negative binomial distribution provides better estimates, even though a sensitivity to return level and data set is identified. The consideration of GCM data permits a strong reduction of uncertainties. The present results support the importance of considering explicitly clustering of losses for an adequate risk assessment for economical applications.

Dimensionality reduction assisted tensor clustering

This paper is concerned with tensor clustering with the assistance of dimensionality reduction approaches. A class of formulation for tensor clustering is introduced based on tensor Tucker decomposition models. In this formulation, an extra tensor mode is formed by a collection of tensors of the same dimensions and then used to assist a Tucker decomposition in order to achieve data dimensionality reduction. We design two types of clustering models for the tensors: PCA Tensor Clustering model and Non-negative Tensor Clustering model, by utilizing different regularizations. The tensor clustering can thus be solved by the optimization method based on the alternative coordinate scheme. Interestingly, our experiments show that the proposed models yield comparable or even better performance compared to most recent clustering algorithms based on matrix factorization.

Serial clustering of extratropical cyclones in a multi-model ensemble of historical and future simulations

This study has investigated serial (temporal) clustering of extra-tropical cyclones simulated by 17 climate models that participated in CMIP5. Clustering was estimated by calculating the dispersion (ratio of variance to mean) of 30 December-February counts of Atlantic storm tracks passing nearby each grid point. Results from single historical simulations of 1975-2005 were compared to those from historical ERA40 reanalyses from 1958-2001 ERA40 and single future model projections of 2069-2099 under the RCP4.5 climate change scenario. Models were generally able to capture the broad features in reanalyses reported previously: underdispersion/regularity (i.e. variance less than mean) in the western core of the Atlantic storm track surrounded by overdispersion/clustering (i.e. variance greater than mean) to the north and south and over western Europe. Regression of counts onto North Atlantic Oscillation (NAO) indices revealed that much of the overdispersion in the historical reanalyses and model simulations can be accounted for by NAO variability. Future changes in dispersion were generally found to be small and not consistent across models. The overdispersion statistic, for any 30 year sample, is prone to large amounts of sampling uncertainty that obscures the climate change signal. For example, the projected increase in dispersion for storm counts near London in the CNRMCM5 model is 0.1 compared to a standard deviation of 0.25. Projected changes in the mean and variance of NAO are insufficient to create changes in overdispersion that are discernible above natural sampling variations.