DATA ANALYTICS 2014, The Third International Conference on Data Analytics
Full texts of third international conference on data analytics are presented.
A common network representation of the stock market is based on correlations of time series of return fluctuations. It is well known that financial time series have a stochastic nature. Therefore there is uncertainty in inference about filtered structures in market network. Thus market network analysis need to be complemented by estimation of uncertainty of the obtained results. However as far as we know there are no relevant research in the literature. In the present paper we maake the first step in this direction. We propose the approach to measure statistical uncertainty of different market network structures. This approach is based on conditional risk for corresponding multiple decision statistical procedures. The proposed appoach is illustrated by numerical evaluation of statistical ucertainty for popular network structures. Our experimental study validates the possibility of application of the approach for comparison of uncerttainty of different network structures.
Similarity searching has a vast range of applications in various fields of computer science. Many methods have been proposed for exact search, but they all suffer from the curse of dimensionality and are, thus, not applicable to high dimensional spaces. Approximate search methods are considerably more efficient in high dimensional spaces. Unfortunately, there are few theoretical results regarding the complexity of these methods and there are no comprehensive empirical evaluations, especially for non-metric spaces. To fill this gap, we present an empirical analysis of data structures for approximate nearest neighbor search in high dimensional spaces. We provide a comparison with recently published algorithms on several data sets. Our results show that small world approaches provide some of the best tradeoffs between efficiency and effectiveness in both metric and non-metric spaces.