Aleksander B. Demko

Mapping high-dimensional data onto a relative distance plane: an exact method for visualizing and characterizing high-dimensional patterns

We introduce a distance (similarity)-based mapping for the visualization of high-dimensional patterns and their relative relationships. The mapping preserves exactly the original distances between points with respect to any two reference patterns in a special two-dimensional coordinate system, the relative distance plane (RDP). As only a single calculation of a distance matrix is required, this method is computationally efficient, an essential requirement for any exploratory data analysis. The data visualization afforded by this representation permits a rapid assessment of class pattern distributions. In particular, we can determine with a simple statistical test whether both training and validation sets of a 2-class, high-dimensional dataset derive from the same class distributions. We can explore any dataset in detail by identifying the subset of reference pairs whose members belong to different classes, cycling through this subset, and for each pair, mapping the remaining patterns. These multiple viewpoints facilitate the identification and confirmation of outliers. We demonstrate the effectiveness of this method on several complex biomedical datasets. Because of its efficiency, effectiveness, and versatility, one may use the RDP representation as an initial, data mining exploration that precedes classification by some classifier. Once final enhancements to the RDP mapping software are completed, we plan to make it freely available to researchers.

Scopira - a system for the analysis of biomedical data

With the proliferation of high-dimensional biomedical data, an acute need exists for a comprehensive, user-friendly software suite that allows investigators, in the health care disciplines, to classify their data through the detection of discriminating features. Scopira is a software initiative that attempts to achieve these goals in addition to providing intuitive visual computation, logic construction and parallel execution. We describe the architecture of Scopira, and various design and implementation issues that surfaced during development.

Scopira: an open source C++ framework for biomedical data analysis applications -- a research project report

In MRI research labs, algorithms are typically implemented in MATLAB or IDL. If performance is an issue they are ported to C and integrated with interpreted systems, not fully utilizing object-oriented software development. This paper presents Scopira, an open source C++ framework suitable for MRI data analysis and visualization.

A Software Development Framework for Agent-Based Infectious Disease Modelling

From the Black Death of 1347–1350 (Murray, 2007) and the Spanish influenza pandemic of 1918–1919 (Taubenberger & Morens, 2006), to the more recent 2003 SARS outbreak (Lingappa et al., 2004) and the 2009 influenza pandemic (Moghadas et al., 2009), as well as countless outbreaks of childhood infections, infectious diseases have been the bane of humanity throughout its existence causing significant morbidity, mortality, and socioeconomic upheaval. Advanced modelling technologies, which incorporate the most current knowledge of virology, immunology, epidemiology, vaccines, antiviral drugs, and public health, have recently come to the fore in identifying effective disease mitigation strategies, and are being increasingly used by public health experts in the study of both epidemiology and pathogenesis. Tracing its historical roots from the pioneering work of Daniel Bernoulli on smallpox (Bernoulli, 1760) to the classical compartmental approach of Kermack and McKendrick (Kermack & McKendrick, 1927), modelling has evolved to deal with data that is more heterogeneous, less coarse (based at a community or individual level), and more complex (joint spatial, temporal and behavioural interactions). This evolution is typified by the agent-based model (ABM) paradigm, lattice-distributed collections of autonomous decision-making entities (agents), the interactions of which unveil the dynamics and emergent properties of the infectious disease outbreak under investigation. The flexibility of ABMs permits an effective representation of the complementary interactions between individuals characterized by localized properties and populations at a global level. However, with flexibility comes complexity; hence, the software implementation of an ABM demands more stringent software design requirements than conventional (and simpler) models of the spread and control of infectious diseases, especially with respect to outcome reproducibility, error detection and system management. Outcome reproducibility is a challenge because emergent properties are not analytically tractable, which is further exacerbated by subtle and difficult to detect errors in algorithm logic and software design. System management of software simulating populations/individuals and biological /physical interactions is a serious challenge, as the implementation will involve distributed (parallelized), non-linear, complex, and multiple processes operating in concert. Given these

Discrimination of software quality in a biomedical data analysis system

Object-oriented visualization-based software systems for biomedical data analysis must deal with complex and voluminous datasets within a flexible yet intuitive graphical user interface. In a research environment, the development of such systems are difficult to manage due to rapidly changing requirements, incorporation of newly developed algorithms, and the needs imposed by a diverse user base. One issue that research supervisors must contend with is an assessment of the quality of the systems software objects with respect to their extensibility, reusability, clarity, and efficiency. Objects from a biomedical data analysis system were independently analyzed by two software architects and ranked according to their quality. Quantitative software features were also compiled at varying levels of granularity. The discriminatory power of these software metrics is discussed and their effectiveness in assessing and predicting software object quality is described.

The Analysis of Software Complexity Using Stochastic Metric Selection

The automated prediction of qualitative attributes such as software complexity is a desirable software engineering goal. A potential technique is to use software metrics as quantitative predictors for these kinds of attributes. We describe a pattern classification method where a large collection of classifiers is presented with randomly selected subsets of software metrics describing modules from a sophisticated biomedical data analysis system. The method identifies the software metric subset that has the highest discriminatory power vis-à-vis software complexity. That is, we identify the metric subset that is most effective at predicting this qualitative attribute. This classification method is empirically evaluated and carefully validated against three benchmark approaches. We demonstrate that this method has utility in the automated prediction of software complexity using quantitative

A Pattern Recognition Application Framework for Biomedical Datasets

: Scopira facilitates the development of high-performance applications by providing many useful subsystems, flexible and efficient data models, low-level tools such as memory management and serialization, GUI constructs, high-level visualization modules, and the ability to implement parallel algorithms with MPI. Scopira plug-in extensions have been developed to enable Matlab scripts to easily call any Scopira module, thus facilitating the migration of prototypes to highly efficient C++ applications. Scopira is continuously under development and future capabilities will include the ability to develop distributed programs using agents, applicable to grid-computing data mining applications. Scopira has proven to be a successful programming framework for implementing high-performance biomedical data analysis applications. It is based on C++, an efficient object-oriented language, and the source code is available as an open-source project for other researchers to use and adapt to their own research endeavours. Scopira has been compiled to work on Linux and Windows XP operating systems with a port to the Mac OS under development. Scopira, EvIdent and RDP are freely available for download from www.scopira.org.

A projection method for the visualization of high-dimensional biomedical datasets

The analysis and interpretation of high-dimensional biomedical datasets for the purposes of confirmatory or exploratory data analysis is a challenging problem. The process raises issues that are not only typically associated with high-dimensional data but also with software implementations of the visualization models. The relative distance plane projection method, which uses a distance-based mapping for visualizing high dimensional patterns and their relative relationships, addresses these confounding factors. This paper describes the algorithm, its implementation in software, and the specialized user interface. Its functionality is demonstrated using a high-dimensional biomedical dataset.

Scopira: a pattern recognition application framework for biomedical datasets

Machine learning techniques are widely used in the analysis of biomedical datasets. Modern devices tend to produce voluminous, high-dimensional datasets for which medical practitioners require high-performance, user-friendly programs and researchers need effective algorithm development and testing platforms. Interactive development systems, such as MATIAB, provide for rapid prototyping of algorithms and visualization but at the cost of computational efficiency. We present Scopira, a C++, open source programming framework for the development of biomedical data analysis applications.

A Classification Canvas for the analysis of biomedical data

With the rapid proliferation of complex high-dimensional biomedical data, an acute need exists for a comprehensive, knowledge-based, domain-specific, user-friendly software suite that allows investigators, in the health care disciplines, to classify their data through the detection of novel or discriminating features therein. The Classification Canvas is an attempt to achieve these goals in addition to providing intuitive visual computation and logic construction. In this paper we describe various design and implementation issues such as: balancing user (novice) friendliness and developer (experienced) utility, performance versus modularity trade-offs, C++ and Java data sharing responsibilities, and creating graphical interfaces for (user-supplied) algorithm control.