Data visualisation and exploration with prior knowledge

Visualising data for exploratory analysis is a major challenge in many applications. Visualisation allows scientists to gain insight into the structure and distribution of the data, for example finding common patterns and relationships between samples as well as variables. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are employed. These methods are favoured because of their simplicity, but they cannot cope with missing data and it is difficult to incorporate prior knowledge about properties of the variable space into the analysis; this is particularly important in the high-dimensional, sparse datasets typical in geochemistry. In this paper we show how to utilise a block-structured correlation matrix using a modification of a well known non-linear probabilistic visualisation model, the Generative Topographic Mapping (GTM), which can cope with missing data. The block structure supports direct modelling of strongly correlated variables. We show that including prior structural information it is possible to improve both the data visualisation and the model fit. These benefits are demonstrated on artificial data as well as a real geochemical dataset used for oil exploration, where the proposed modifications improved the missing data imputation results by 3 to 13%.

β-amyloid (Aβ) deposition in cognitively normal brain, dementia with Lewy bodies, and Alzheimer's disease:a study using principal components analysis

The densities of diffuse, primitive, and classic ß-amyloid (Aß) deposits were studied in the temporal lobe in cognitively normal brain, dementia with Lewy bodies (DLB), familial Alzheimer’s disease (FAD), and sporadic AD (SAD). Principal components analysis (PCA) was used to determine whether there were distinct differences between groups or whether Aß pathology was more continuously distributed from group to group. Three principal components (PC) were extracted from the data accounting for 56% of the total variance. Plots of cases in relation to the PC did not result in distinct groups but suggested overlap in Aß deposition between the groups. In addition, there were linear correlations between the densities of Aß deposits and the distribution of the cases along the PC in specific brain regions suggesting continuous variation from group to group. PC1 was associated with the degree of maturation of Aß deposits, PC2 with differences between FAD and SAD, and PC3 with the degree of spread of Aß pathology into the hippocampus. Apolipoprotein E (APOE) genotype was not associated with variation in Aß deposition between cases. PCA may be a useful method of studying the pathological interface between closely related neurodegenerative disorders.

Understanding U.S. regional linguistic variation with Twitter data analysis

We analyze a Big Data set of geo-tagged tweets for a year (Oct. 2013–Oct. 2014) to understand the regional linguistic variation in the U.S. Prior work on regional linguistic variations usually took a long time to collect data and focused on either rural or urban areas. Geo-tagged Twitter data offers an unprecedented database with rich linguistic representation of fine spatiotemporal resolution and continuity. From the one-year Twitter corpus, we extract lexical characteristics for twitter users by summarizing the frequencies of a set of lexical alternations that each user has used. We spatially aggregate and smooth each lexical characteristic to derive county-based linguistic variables, from which orthogonal dimensions are extracted using the principal component analysis (PCA). Finally a regionalization method is used to discover hierarchical dialect regions using the PCA components. The regionalization results reveal interesting linguistic regional variations in the U.S. The discovered regions not only confirm past research findings in the literature but also provide new insights and a more detailed understanding of very recent linguistic patterns in the U.S.

Data visualization during the early stages of drug discovery

Multidimensional compound optimization is a new paradigm in the drug discovery process, yielding efficiencies during early stages and reducing attrition in the later stages of drug development. The success of this strategy relies heavily on understanding this multidimensional data and extracting useful information from it. This paper demonstrates how principled visualization algorithms can be used to understand and explore a large data set created in the early stages of drug discovery. The experiments presented are performed on a real-world data set comprising biological activity data and some whole-molecular physicochemical properties. Data visualization is a popular way of presenting complex data in a simpler form. We have applied powerful principled visualization methods, such as generative topographic mapping (GTM) and hierarchical GTM (HGTM), to help the domain experts (screening scientists, chemists, biologists, etc.) understand and draw meaningful decisions. We also benchmark these principled methods against relatively better known visualization approaches, principal component analysis (PCA), Sammon's mapping, and self-organizing maps (SOMs), to demonstrate their enhanced power to help the user visualize the large multidimensional data sets one has to deal with during the early stages of the drug discovery process. The results reported clearly show that the GTM and HGTM algorithms allow the user to cluster active compounds for different targets and understand them better than the benchmarks. An interactive software tool supporting these visualization algorithms was provided to the domain experts. The tool facilitates the domain experts by exploration of the projection obtained from the visualization algorithms providing facilities such as parallel coordinate plots, magnification factors, directional curvatures, and integration with industry standard software. © 2006 American Chemical Society.

Independent component analysis for domain independent watermarking

A new principled domain independent watermarking framework is presented. The new approach is based on embedding the message in statistically independent sources of the covertext to mimimise covertext distortion, maximise the information embedding rate and improve the method's robustness against various attacks. Experiments comparing the performance of the new approach, on several standard attacks show the current proposed approach to be competitive with other state of the art domain-specific methods.

Watermarking of audio signals using independent component analysis

A novel approach to watermarking of audio signals using Independent Component Analysis (ICA) is proposed. It exploits the statistical independence of components obtained by practical ICA algorithms to provide a robust watermarking scheme with high information rate and low distortion. Numerical simulations have been performed on audio signals, showing good robustness of the watermark against common attacks with unnoticeable distortion, even for high information rates. An important aspect of the method is its domain independence: it can be used to hide information in other types of data, with minor technical adaptations.

miniDVMS v1.8 : A user manual(The data visualization and modeling system)

Today, the data available to tackle many scientific challenges is vast in quantity and diverse in nature. The exploration of heterogeneous information spaces requires suitable mining algorithms as well as effective visual interfaces. miniDVMS v1.8 provides a flexible visual data mining framework which combines advanced projection algorithms developed in the machine learning domain and visual techniques developed in the information visualisation domain. The advantage of this interface is that the user is directly involved in the data mining process. Principled projection methods, such as generative topographic mapping (GTM) and hierarchical GTM (HGTM), are integrated with powerful visual techniques, such as magnification factors, directional curvatures, parallel coordinates, and user interaction facilities, to provide this integrated visual data mining framework. The software also supports conventional visualisation techniques such as principal component analysis (PCA), Neuroscale, and PhiVis. This user manual gives an overview of the purpose of the software tool, highlights some of the issues to be taken care while creating a new model, and provides information about how to install and use the tool. The user manual does not require the readers to have familiarity with the algorithms it implements. Basic computing skills are enough to operate the software.

Principal components analysis of Alzheimer’s disease based on neuropathological data: A study of 79 patients

A Principal Components Analysis of neuropathological data from 79 Alzheimer’s disease (AD) cases was performed to determine whether there was evidence for subtypes of the disease. Two principal components were extracted from the data which accounted for 72% and 12% of the total variance respectively. The results suggested that 1) AD was heterogeneous but subtypes could not be clearly defined; 2) the heterogeneity, in part, reflected disease onset; 3) familial cases did not constitute a distinct subtype of AD and 4) there were two forms of late onset AD, one of which was associated with less senile plaque and neurofibrillary tangle development but with a greater degree of brain atherosclerosis.

Visualization of molecular fingerprints

A visualization plot of a data set of molecular data is a useful tool for gaining insight into a set of molecules. In chemoinformatics, most visualization plots are of molecular descriptors, and the statistical model most often used to produce a visualization is principal component analysis (PCA). This paper takes PCA, together with four other statistical models (NeuroScale, GTM, LTM, and LTM-LIN), and evaluates their ability to produce clustering in visualizations not of molecular descriptors but of molecular fingerprints. Two different tasks are addressed: understanding structural information (particularly combinatorial libraries) and relating structure to activity. The quality of the visualizations is compared both subjectively (by visual inspection) and objectively (with global distance comparisons and local k-nearest-neighbor predictors). On the data sets used to evaluate clustering by structure, LTM is found to perform significantly better than the other models. In particular, the clusters in LTM visualization space are consistent with the relationships between the core scaffolds that define the combinatorial sublibraries. On the data sets used to evaluate clustering by activity, LTM again gives the best performance but by a smaller margin. The results of this paper demonstrate the value of using both a nonlinear projection map and a Bernoulli noise model for modeling binary data.

Novel visualization methods for protein data

Visualization of high-dimensional data has always been a challenging task. Here we discuss and propose variants of non-linear data projection methods (Generative Topographic Mapping (GTM) and GTM with simultaneous feature saliency (GTM-FS)) that are adapted to be effective on very high-dimensional data. The adaptations use log space values at certain steps of the Expectation Maximization (EM) algorithm and during the visualization process. We have tested the proposed algorithms by visualizing electrostatic potential data for Major Histocompatibility Complex (MHC) class-I proteins. The experiments show that the variation in the original version of GTM and GTM-FS worked successfully with data of more than 2000 dimensions and we compare the results with other linear/nonlinear projection methods: Principal Component Analysis (PCA), Neuroscale (NSC) and Gaussian Process Latent Variable Model (GPLVM).

Principal components analysis of Alzheimer's disease based on neuropathological data: a study of 79 patients

A principal components analysis was carried out on neuropathological data collected from 79 cases of Alzheimer's disease (AD) diagnosed in a single centre. The purpose of the study was to determine whether on neuropathological criteria there was evidence for clearly defined subtypes of the disease. Two principal components (PC1 and PC2) were extracted from the data. PC1 was considerable more important than PC2 accounting for 72% of the total variance. When plotted in relation to the first two principal components the majority of cases (65/79) were distributed in a single cluster within which subgroupings were not clearly evident. In addition, there were a number of individual, mainly early-onset cases, which were neither related to each other nor to the main cluster. The distribution of each neuropathological feature was examined in relation to PC1 and 2, Disease onset, rhe degree of gross brain atrophy, neuronal loss and the devlopment of senile plaques (SP) and neurofibrillary tangles (NFT) were negatively correlated with PC1. The devlopment of SP and NFT and the degree of brain athersclerosis were positively correlated with PC2. These results suggested: 1) that there were different forms of AD but no clear division of the cases into subclasses could be made based on the neuropathological criteria used; the cases showing a more continuous distribution from one form to another, 2) that disease onset was an important variable and was associated with a greater development of pathological changes, 3) familial cases were not a distinct subclass of AD; the cases being widely distributed in relation to PC1 and PC2 and 4) that there may be two forms of late-onset AD whic grade into each other, one of which was associated with less SP and NFT development but with a greater degree of brain atherosclerosis.

A novel detection method for determining a surgeon's fatigue and hand tremor

The thesis presents new methodology and algorithms that can be used to analyse and measure the hand tremor and fatigue of surgeons while performing surgery. This will assist them in deriving useful information about their fatigue levels, and make them aware of the changes in their tool point accuracies. This thesis proposes that muscular changes of surgeons, which occur through a day of operating, can be monitored using Electromyography (EMG) signals. The multi-channel EMG signals are measured at different muscles in the upper arm of surgeons. The dependence of EMG signals has been examined to test the hypothesis that EMG signals are coupled with and dependent on each other. The results demonstrated that EMG signals collected from different channels while mimicking an operating posture are independent. Consequently, single channel fatigue analysis has been performed. In measuring hand tremor, a new method for determining the maximum tremor amplitude using Principal Component Analysis (PCA) and a new technique to detrend acceleration signals using Empirical Mode Decomposition algorithm were introduced. This tremor determination method is more representative for surgeons and it is suggested as an alternative fatigue measure. This was combined with the complexity analysis method, and applied to surgically captured data to determine if operating has an effect on a surgeon’s fatigue and tremor levels. It was found that surgical tremor and fatigue are developed throughout a day of operating and that this could be determined based solely on their initial values. Finally, several Nonlinear AutoRegressive with eXogenous inputs (NARX) neural networks were evaluated. The results suggest that it is possible to monitor surgeon tremor variations during surgery from their EMG fatigue measurements.

Uses of frontier efficiency methodologies and multi-criteria decision making for performance measurement in the energy sector

Guest editorial Ali Emrouznejad is a Senior Lecturer at the Aston Business School in Birmingham, UK. His areas of research interest include performance measurement and management, efficiency and productivity analysis as well as data mining. He has published widely in various international journals. He is an Associate Editor of IMA Journal of Management Mathematics and Guest Editor to several special issues of journals including Journal of Operational Research Society, Annals of Operations Research, Journal of Medical Systems, and International Journal of Energy Management Sector. He is in the editorial board of several international journals and co-founder of Performance Improvement Management Software. William Ho is a Senior Lecturer at the Aston University Business School. Before joining Aston in 2005, he had worked as a Research Associate in the Department of Industrial and Systems Engineering at the Hong Kong Polytechnic University. His research interests include supply chain management, production and operations management, and operations research. He has published extensively in various international journals like Computers & Operations Research, Engineering Applications of Artificial Intelligence, European Journal of Operational Research, Expert Systems with Applications, International Journal of Production Economics, International Journal of Production Research, Supply Chain Management: An International Journal, and so on. His first authored book was published in 2006. He is an Editorial Board member of the International Journal of Advanced Manufacturing Technology and an Associate Editor of the OR Insight Journal. Currently, he is a Scholar of the Advanced Institute of Management Research. Uses of frontier efficiency methodologies and multi-criteria decision making for performance measurement in the energy sector This special issue aims to focus on holistic, applied research on performance measurement in energy sector management and for publication of relevant applied research to bridge the gap between industry and academia. After a rigorous refereeing process, seven papers were included in this special issue. The volume opens with five data envelopment analysis (DEA)-based papers. Wu et al. apply the DEA-based Malmquist index to evaluate the changes in relative efficiency and the total factor productivity of coal-fired electricity generation of 30 Chinese administrative regions from 1999 to 2007. Factors considered in the model include fuel consumption, labor, capital, sulphur dioxide emissions, and electricity generated. The authors reveal that the east provinces were relatively and technically more efficient, whereas the west provinces had the highest growth rate in the period studied. Ioannis E. Tsolas applies the DEA approach to assess the performance of Greek fossil fuel-fired power stations taking undesirable outputs into consideration, such as carbon dioxide and sulphur dioxide emissions. In addition, the bootstrapping approach is deployed to address the uncertainty surrounding DEA point estimates, and provide bias-corrected estimations and confidence intervals for the point estimates. The author revealed from the sample that the non-lignite-fired stations are on an average more efficient than the lignite-fired stations. Maethee Mekaroonreung and Andrew L. Johnson compare the relative performance of three DEA-based measures, which estimate production frontiers and evaluate the relative efficiency of 113 US petroleum refineries while considering undesirable outputs. Three inputs (capital, energy consumption, and crude oil consumption), two desirable outputs (gasoline and distillate generation), and an undesirable output (toxic release) are considered in the DEA models. The authors discover that refineries in the Rocky Mountain region performed the best, and about 60 percent of oil refineries in the sample could improve their efficiencies further. H. Omrani, A. Azadeh, S. F. Ghaderi, and S. Abdollahzadeh presented an integrated approach, combining DEA, corrected ordinary least squares (COLS), and principal component analysis (PCA) methods, to calculate the relative efficiency scores of 26 Iranian electricity distribution units from 2003 to 2006. Specifically, both DEA and COLS are used to check three internal consistency conditions, whereas PCA is used to verify and validate the final ranking results of either DEA (consistency) or DEA-COLS (non-consistency). Three inputs (network length, transformer capacity, and number of employees) and two outputs (number of customers and total electricity sales) are considered in the model. Virendra Ajodhia applied three DEA-based models to evaluate the relative performance of 20 electricity distribution firms from the UK and the Netherlands. The first model is a traditional DEA model for analyzing cost-only efficiency. The second model includes (inverse) quality by modelling total customer minutes lost as an input data. The third model is based on the idea of using total social costs, including the firm’s private costs and the interruption costs incurred by consumers, as an input. Both energy-delivered and number of consumers are treated as the outputs in the models. After five DEA papers, Stelios Grafakos, Alexandros Flamos, Vlasis Oikonomou, and D. Zevgolis presented a multiple criteria analysis weighting approach to evaluate the energy and climate policy. The proposed approach is akin to the analytic hierarchy process, which consists of pairwise comparisons, consistency verification, and criteria prioritization. In the approach, stakeholders and experts in the energy policy field are incorporated in the evaluation process by providing an interactive mean with verbal, numerical, and visual representation of their preferences. A total of 14 evaluation criteria were considered and classified into four objectives, such as climate change mitigation, energy effectiveness, socioeconomic, and competitiveness and technology. Finally, Borge Hess applied the stochastic frontier analysis approach to analyze the impact of various business strategies, including acquisition, holding structures, and joint ventures, on a firm’s efficiency within a sample of 47 natural gas transmission pipelines in the USA from 1996 to 2005. The author finds that there were no significant changes in the firm’s efficiency by an acquisition, and there is a weak evidence for efficiency improvements caused by the new shareholder. Besides, the author discovers that parent companies appear not to influence a subsidiary’s efficiency positively. In addition, the analysis shows a negative impact of a joint venture on technical efficiency of the pipeline company. To conclude, we are grateful to all the authors for their contribution, and all the reviewers for their constructive comments, which made this special issue possible. We hope that this issue would contribute significantly to performance improvement of the energy sector.

Independent component analysis and feature extraction of financial time series

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

Chemometric modelling to relate antioxidants, neutral lipid fatty acids and flavour components in chicken breast

Relationships among quality factors in retailed free-range, corn-fed, organic, and conventional chicken breasts (9) were modeled using chemometric approaches. Use of principal component analysis (PCA) to neutral lipid composition data explained the majority (93%) of variability (variance) in fatty acid contents in 2 significant multivariate factors. PCA explained 88 and 75% variance in 3 factors for, respectively, flame ionization detection (FID) and nitrogen phosphorus (NPD) components in chromatographic flavor data from cooked chicken after simultaneous distillation extraction. Relationships to tissue antioxidant contents were modeled. Partial least square regression (PLS2), interrelating total data matrices, provided no useful models. By using single antioxidants as Y variables in PLS (1), good models (r2 values > 0.9) were obtained for alpha-tocopherol, glutathione, catalase, glutathione peroxidase, and reductase and FID flavor components and among the variables total mono and polyunsaturated fatty acids and subsets of FID, and saturated fatty acid and NPD components. Alpha-tocopherol had a modest (r2 = 0.63) relationship with neutral lipid n-3 fatty acid content. Such factors thus relate to flavor development and quality in chicken breast meat.