Resulting Cluster Structure Research Articles

Identifying subtypes of mild cognitive impairment in Parkinson's disease using cluster analysis.

The concept of Mild Cognitive Impairment (MCI) in Parkinson's disease (PD) has shown the potential for identifying at-risk dementia patients. Identifying subtypes of MCI is likely to assist therapeutic discoveries and better clinical management of patients with PD (PWP). Recent cluster-based approaches have demonstrated dominance in memory and executive impairment in PD. The present study will further explore the role of memory and executive impairment and associated clinical features in non-demented PWP. A K-means cluster analysis was performed on ten "frontal" and "posterior" cognitive variables derived from a dataset of 85 non-demented PWP. The resulting cluster structure was chosen based on quantitative, qualitative, theoretical, and clinical validity. Cluster profiles were then created through statistical analysis of cognitive and clinical/demographic variables. A descriptive analysis of each cluster's performance on a comprehensive PD-MCI diagnostic battery was also explored. The resulting cluster structure revealed four distinct cognitive phenotypes: (1) frontal-dominant impairment; (2) posterior-cortical-dominant impairment; (3) global impairment, and (4) cognitively intact. Demographic profiling revealed significant differences in the age, gender split, global cognitive ability, and motor symptoms between these clusters. However, there were no significant differences between the clusters on measures of depression, apathy, and anxiety. These results validate the existence of distinct cognitive phenotypes within PD-MCI and encourage future research into their clinical trajectory and neuroimaging correlates.

Genie+OWA: Robustifying hierarchical clustering with OWA-based linkages

We investigate the application of the Ordered Weighted Averaging (OWA) data fusion operator in agglomerative hierarchical clustering. The examined setting generalises the well-known single, complete and average linkage schemes. It allows to embody expert knowledge in the cluster merge process and to provide a much wider range of possible linkages. We analyse various families of weighting functions on numerous benchmark data sets in order to assess their influence on the resulting cluster structure. Moreover, we inspect the correction for the inequality of cluster size distribution – similar to the one in the Genie algorithm. Our results demonstrate that by robustifying the procedure with the Genie correction, we can obtain a significant performance boost in terms of clustering quality. This is particularly beneficial in the case of the linkages based on the closest distances between clusters, including the single linkage and its “smoothed” counterparts. To explain this behaviour, we propose a new linkage process called three-stage OWA which yields further improvements. This way we confirm the intuition that hierarchical cluster analysis should rather take into account a few nearest neighbours of each point, instead of trying to adapt to their non-local neighbourhood.

Towards the crystal structure of thymine: An intermolecular force field development and parallel global cluster optimizations.

A new ab initio potential for the thymine dimer has been developed by using a combination of density functional theory and symmetry-adapted perturbation theory (DFT-SAPT). It has been shown that DFT-SAPT is in very good agreement with the reference CCSD(T) calculations performed for the hydrogen-bonded and stacked thymine dimers. Parallel global cluster optimizations have been carried out employing the new force field from the dimer to n = 50. It has been observed that the resulting cluster structures tend to form the building blocks of the crystal structure of thymine. This study therefore highlights the importance of dimer potentials to gain insights about the crystal structures.

Partitioning of polycyclic aromatic hydrocarbons in heterogeneous clusters

The morphologies of heterogeneous clusters of polycyclic aromatic hydrocarbons (PAHs) are investigated using molecular modelling. Clusters of up to 100 molecules containing combinations of the different sized PAHs circumcoronene, coronene, ovalene, or pyrene are evaluated. Replica exchange molecular dynamics simulations using an all-atom force field parameterised for PAHs sample many configurations at high and low temperatures to determine stable low energy structures. The resulting cluster structures are evaluated using molecular radial distances and coordination numbers, and are found to be independent of initial configuration and the cluster sizes studied. Stable clusters consist of stacked PAHs in a core-shell structure, where the larger PAHs are found closer to the cluster core and the smaller PAHs are located on the cluster surface. This work provides novel insight into the molecular partitioning of heterogeneous aromatic clusters, with particular relevance to the structure of nascent soot particles.

Tipology of Value Orientations of Ukrainians (Based on the Data of European Social Survey)

The article is devoted to the analysis of value orientations of Ukrainians based on the data of six waves of the European Social Survey (2004-2012). The aim is to identify the structure of Ukrainian society on the basis of the value orientations of respondents, obtained by the public opinion polling. The dynamics of this structure is also analyzed during 2004-2012. The concept of "value" in this article is defined within Schwartz's theory as desirable goals that go beyond specific situations, differ in importance from each other and are guiding principles in human life. The analysis uses the ten Schwarz's values, which are grouped into four value sectors, which are named «Conservation», «Self-Enhancement», «Self-Transcendence» and «Openness to Change». To identify the structure respondents were divided into three clusters, depending on the expressiveness of value orientations in each sector. The clustering was carried out using the k-means method, while the cluster centers were wrote based on the theoretical conception: the first cluster includes respondents with the most expressed values of «Self-Enhancement» and «Openness to Change», the second one – respondents with the most expressed values of «Self-Transcendence» and «Conservation», the third one – respondents who are closes to average values of all four sectors. The resulting cluster structure was identified in each of the six waves of the European Social Survey. At the same time clusters are differ in a number of socio-demographic indicators, as well as in some social orientations. The third cluster is biggest in the each wave of the survey. It means that the group of respondent who doesn’t have strongly expressed values of any of four sectors is the biggest. It can be explained by the phenomenon of ambivalence of the individual that means combination of views, thoughts, feelings, orientations that are mutually exclusive. It was also found that the trend of increasing the expressiveness of the values of «Self-Enhancement» and «Openness to Change», which manifests itself in the array as a whole, is also observed within the first and third clusters, but the value orientations of the second cluster remain relatively stable.

Classification of Holstein Dairy Cattles in Terms of Parameters Some Milk Component Belongs by Using The Fuzzy Cluster Analysis

This study was carried out on classification of Holstein Friesian breed dairy cattles in terms of some milk component parameters and on investigating the relevant parameters in the resulting cluster structures. Within the scope of this study, thirteen different criteria were used including somatic cell count (SCC), milk fat (%), milk protein (%), milk lactose (%), casein (%), urea (%), dry matter (%), non-fat dry matter (%), density (g/cm3), acidity (oSH), free fatty acids (mmol/10L), citric acid (%) and freezing point (oC). As a result of the analysis using Fanny algorithm based on the principle of fuzzy equality, the fuzziness level was found to be minimum when a total of 282 cattles were divided into 2 clusters with the accuracy rate of 97.5%. Accordingly, the cattles were classified in terms of the investigated characteristics in 2 different clusters in which 25 cattles were in Cluster 1 and the rest of the cattles were in Cluster 2. When the resulting cluster structures were studied, it was found that Cluster 2 has a more stable clustering than Cluster 1. When evaluating the change in milk components according to the clusters, it was concluded that somatic cell count, dry matter (%), milk fat (%) and density (g/cm3) have significant differences between clusters (P 0.05).

Paving the way for next generation data-stream clustering: towards a unique and statistically valid cluster structure at any time step

In the domain of data-stream clustering, e.g., dynamic text mining as our application domain, our goal is two-fold and a long term one: 1 at each data input, the resulting cluster structure has to be unique, independent of the order the input vectors are presented 2 this structure has to be meaningful for an expert, e.g., not composed of a huge 'catch-all' cluster in a cloud of tiny specific ones, as is often the case with large sparse data tables. The first preliminary condition is satisfied by our Germen density-mode seeking algorithm, but the relevance of the clusters vis-a-vis expert judgment relies on the definition of a data density, relying itself on the type of graph chosen for embedding the similarities between text inputs. Having already demonstrated the dynamic behaviour of Germen algorithm, we focus here on appending a Monte-Carlo method for extracting statistically valid inter-text links, which looks promising applied both to an excerpt of the Pascal bibliographic database, and to the Reuters-RCV1 news test collection. Though not being a central issue here, the time complexity of our algorithms is eventually discussed.

High-salt stabilization of Laponite clay particles

Colloidal radioactive transuranic wastes are currently buried in large tanks in the form of dense colloids in high salt, high pH aqueous media. These facilities are beginning to fail, so it is necessary to transport and “package” them for more permanent disposal, processes requiring understanding of the microstructure that develops under such conditions. Laponite RD clay is believed to be a good simulant for the colloids in the waste tanks, and the present study addresses their behavior under high salt conditions, where previous studies have frequently observed the phenomenon of “restabilization,” i.e., the attainment of aggregation stability at high electrolyte conditions. Specifically, the aggregation kinetics and the resulting cluster structure (fractal dimension) of Laponite RD clay colloids at high concentrations of BaCl 2 (an electrolyte previously shown to lead to restabilization) are investigated. At low-to-intermediate electrolyte concentrations, the clay is found to behave in accord with DLVO theory, i.e., low salt conditions yield slow aggregation into densely-packed aggregates, whereas intermediate salt concentrations, sufficient to cause double layer collapse, produce rapid aggregation into open aggregates. High salt concentrations, however, show slow rates of aggregation. The aggregate structure under these conditions is found to mimic that found for very low electrolyte concentrations, i.e., high fractal dimension. Further experiments show that a sudden increase in salt concentration in a system containing young open aggregates produced under intermediate salt concentrations causes them to reform into more compact structures.

A access-based clustering protocol for multihop wireless ad hoc networks

The ad hoc network is a temporary wireless system without a fixed (wired or wireless) infrastructure. Many clustering algorithms have been proposed to partition mobile users into clusters to support routing and network management. Most previous studies, however, focus on algorithm design, lacking an overall evaluation of clustering overheads. We design a multiple access scheme for the broadcast of control messages, and propose a new access-based clustering protocol (ABCP) whose cluster formation is heavily influenced by the outcome of the multiple access. By taking into account many delicate aspects in the clustering process, the ABCP provides a generic, flexible, rapidly deployed and stable cluster architecture for the upper layer protocols. Simulation is used to compare ABCP with the other clustering strategy in terms of cluster stability and overheads. Since ABCP makes clustering decision directly based on the result of channel access, it requires fewer control overheads and has shorter convergence time than the other clustering criteria. We also demonstrate that the resulting cluster structure by ABCP behaves stable in face of topology changes.

Generality-based conceptual clustering with probabilistic concepts

Statistical research in clustering has almost universally focused on data sets described by continuous features and its methods are difficult to apply to tasks involving symbolic features. In addition, these methods are seldom concerned with helping the user in interpreting the results obtained. Machine learning researchers have developed conceptual clustering methods aimed at solving these problems. Following a long term tradition in AI, early conceptual clustering implementations employed logic as the mechanism of concept representation. However, logical representations have been criticized for constraining the resulting cluster structures to be described by necessary and sufficient conditions. An alternative are probabilistic concepts which associate a probability or weight with each property of the concept definition. In this paper, we propose a symbolic hierarchical clustering model that makes use of probabilistic representations and extends the traditional ideas of specificity-generality typically found in machine learning. We propose a parameterized measure that allows users to specify both the number of levels and the degree of generality of each level. By providing some feedback to the user about the balance of the generality of the concepts created at each level and given the intuitive behavior of the user parameter, the system improves user interaction in the clustering process.

The use of linked line segments for cluster representation and data reduction

Regardless of how one searches for clusters in data, the resulting cluster structures are often modelled as their respective centroids. However, the use of centroids in this manner implies the clusters are hyperspherical in shape. If the cluster shape is not hyperspherical, using the distance between individual observations and the centroid as a metric of an observation's cluster membership can be misleading. This report proposes the use of a linked line segment based model of cluster structure which is not biased towards any particular cluster shape. The effectiveness of the linked line segment approach is demonstrated in a data reduction exercise using simulated and real world data.

Cluster/dislocation interactions in dilute aluminum-based solid solutions

The influence of single solute atoms and solute clusters on an extended edge dislocation dipole in Al was studied by atomistic simulation. Single Cu and Ag solute/dislocation interaction energy calculations showed that Cu interacts strongly with an Al extended dislocation and prefers sites in the compressive region, in agreement with elasticity theory predictions. Single Ag atoms, however, are strongly repelled by an Al extended dislocation, in contrast with elasticity theory predictions. Monte Carlo simulations of Al: 1% Cu, Al: 2% Cu, Al: 1% Ag, Al: 0.5% Cu, 0.5% Ag, and Al: 0.75% Cu, 0.25% Ag were carried out in the presence of an extended dislocation dipole at 600 K allowing for solute segregation. Cu atoms in the binary alloys were observed to segregate to the compressive regions of the extended dislocation dipole, forming widespread “atmospheres” over the width of both extended dislocations which did not affect the partial dislocation spacing. Ag in the binary alloy formed small Ag zones which also had little influence on the spacing between the partials. The ternary systems, however, exhibited highly localized solute clusters that had a large impact on the extended dislocation dipole structure, increasing the separation between the partial dislocations. The resulting cluster structures are discussed along with their influence on the apparent stacking fault energy of the alloy systems.

Cluster analysis of the WAIS on brain-damaged patients.

In recognition of increased usage of cluster analytic techniques, the current study attempted to examine sampling bounds for tests and subjects on one cluster-analytic procedure. Specifically, the WAIS subtest scores from a heterogeneous group of subjects with CNS involvement were first, factor analyzed and then clustered by means of Ward's hierarchical procedure. This methodology had been previously applied to two different samples of aphasics and aphasia batteries. In both these aphasic studies, an underlying severity dimension of language dysfunction was found. In contrast, in the current study, salient feature groups were found. Moreover, these groups were distributed in a sequential or hierarchical model in terms of specific cognitive deficits. From these data, it is argued that two major parameters in determining the resulting cluster structures are the homogeneity/heterogeneity of the sample and test battery employed. Although such a statement is obvious, the theoretical ramifications are large, in that the cluster structure may be a function of test weighting or sample diversity, not the phenomenon of interest. However, the cluster structures found in the current study were consistent with historical models of cognitive dysfunction and also lend support to hierarchical models of sequential cognitive impairment. These findings are discussed in terms of their clinical and experimental ramifications with respect to grouping subjects in terms of WAIS profiles.