A Taxonomy of Big Data for Optimal Predictive Machine Learning and Data Mining

Big data comes in various ways, types, shapes, forms and sizes. Indeed, almost all areas of science, technology, medicine, public health, economics, business, linguistics and social science are bombarded by ever increasing flows of data begging to be analyzed efficiently and effectively. In this paper, we propose a rough idea of a possible taxonomy of big data, along with some of the most commonly used tools for handling each particular category of bigness. The dimensionality p of the input space and the sample size n are usually the main ingredients in the characterization of data bigness. The specific statistical machine learning technique used to handle a particular big data set will depend on which category it falls in within the bigness taxonomy. Large p small n data sets for instance require a different set of tools from the large n small p variety. Among other tools, we discuss Preprocessing, Standardization, Imputation, Projection, Regularization, Penalization, Compression, Reduction, Selection, Kernelization, Hybridization, Parallelization, Aggregation, Randomization, Replication, Sequentialization. Indeed, it is important to emphasize right away that the so-called no free lunch theorem applies here, in the sense that there is no universally superior method that outperforms all other methods on all categories of bigness. It is also important to stress the fact that simplicity in the sense of Ockham’s razor non-plurality principle of parsimony tends to reign supreme when it comes to massive data. We conclude with a comparison of the predictive performance of some of the most commonly used methods on a few data sets.

A Comparative Analysis of Predictive Learning Algorithms on High-Dimensional Microarray Cancer Data

This research evaluates pattern recognition techniques on a subclass of big data where the dimensionality of the input space (p) is much larger than the number of observations (n). Specifically, we evaluate massive gene expression microarray cancer data where the ratio κ is less than one. We explore the statistical and computational challenges inherent in these high dimensional low sample size (HDLSS) problems and present statistical machine learning methods used to tackle and circumvent these difficulties. Regularization and kernel algorithms were explored in this research using seven datasets where κ < 1. These techniques require special attention to tuning necessitating several extensions of cross-validation to be investigated to support better predictive performance. While no single algorithm was universally the best predictor, the regularization technique produced lower test errors in five of the seven datasets studied.

Оптимизация Оценки Вероятности Ошибочной Классификации в Дискретном Случае

* Работа выполнена при поддержке РФФИ, гранты 07-01-00331-a и 08-01-00944-a

Interactive 3D Visualization of Bézier Curves using Java Open Graphics Library (JOGL)

We present a new program tool for interactive 3D visualization of some fundamental algorithms for representation and manipulation of Bézier curves. The program tool has an option for demonstration of one of their most important applications - in graphic design for creating letters by means of cubic Bézier curves. We use Java applet and JOGL as our main visualization techniques. This choice ensures the platform independency of the created applet and contributes to the realistic 3D visualization. The applet provides basic knowledge on the Bézier curves and is appropriate for illustrative and educational purposes. Experimental results are included.

3D Reconstruction and Digital Visualization of the South of the Royal Palace in Great Preslav

The report presents the film 10th century. The South of the Royal Palace in Great Preslav. It consists of two parts – 10th century. The Royal Palace in Great Prelsav. The Square with the Pinnacle and The Ruler’s Lodgings. 3D and virtual reconstructions of an architectural ensemble – part of the Preslav Royal Court unearthed during archaeological researches are used in the film. 3D documentaries have already gained popularity around the world and are well received by both scholars and the public at large. One of the distinguished tourist destinations in Bulgaria is Great Preslav – capital of the mediaeval Bulgarian state and a significant cultural center of the European Southeast in 9th–10th centuries, too. The first part of the film is created with the financial support of America for Bulgaria Foundation and the second – with the funding of Bulgarian National Science Fund at the Ministry of Education, Youth and Science. A team of almost 20 members worked on the film, including computer specialists, professional actors, and translators in the four main European languages – English, German, French and Russian, Trima Sound Recording Studio. In the first part of the 3D film are shown a segment of the Royal Palace, the square with the water pinnacle and the adjacent buildings – an important structural element of the town-planning of the Preslav Court center in the 10th century. In the second part the accent is the southern part of the Royal Palace in Great Preslav, where the personal residence of the Preslav ruler’s dynasty is situated. The work on the virtual reconstruction was done by Virtual Archaeology club at the Mathematical School, Shumen. Due to the efforts of its members it is now clear how the square in front of the southern gate looked like.

Interactive Visualization of the Building of University of Economics – Varna via 3D Modeling

The object of this paper is presenting the University of Economics – Varna, using a 3D model with 3Ds MAX. Created in 1920, May 14, University of Economics - Varna is a cultural institution with a place and style of its own. With the emergence of the three-dimensional modeling we entered a new stage of the evolution of computer graphics. The main target is to preserve the historical vision, to demonstrate forward-thinking and using of future-oriented approaches.