Showing papers by "SITE Santa Fe published in 2011"

Hybrid Data Reduction Technique for Classification of Transaction Data

Abstract: Data classification problems during the process of mining transaction data requires robust and efficient data reduction technique to guard against loss of essential level information. In this paper, we have addressed the concepts of data reduction in transaction processing systems. The tradeoffs of data reduction techniques are being presented and a hybrid technique for data reduction suitable for addressing classification problems of transaction data is proposed.

Competitive Intelligence and Decision Problems

Abstract: The development of a research, teaching, or application of competitive (economic) intelligence requires a strategic and transverse visionin regards torelated issues. It is essential to integrate the role of culturewhen interpreting results, either from the training of a specialist or in respect to a country or region. The authors of this book, members of an expert group supported by the CNRS in France, bring all of their talents together to create a comprehensive book that does just this and more.

Signal Processing Approach for Prediction of Kink in Transmembrane α-Helices

Abstract: The functions of transmembrane proteins are attributed by kinks (bends) in helices. Kinked helices are believed to be required for appropriate helix-helix and protein-protein interaction in membrane protein complexes. Therefore, knowledge of kink and its prediction from amino acid sequences is of great help in understanding the function of proteins. However, determination of kink in transmembrane α-helices is a computationally intensive task. In this paper we have developed signal processing algorithms based on discrete Fourier transform and wavelet transform for prediction of kink in the helices with a prediction efficiency of ~80%. The numerical representation of the protein in terms of probability of occurrence of amino acids constituted in kinked helices contains most of the necessary information in determining the kink location, and the signal processing methods capture this information more effectively than existing statistical and machine learning methods.

Information display from board wargame for marketing strategy identification

Abstract: Marketing warfare is an alternative solution for a company to defend itself or to win market parts. This approach presents consumer spirit as a battleground where companies make military maneuvers to confront each other. But a problem subsists, how make a link between market and battle or war? May be a solution exists: business wargames. But now, they are too complex or only role playing oriented without any solution to map battles. However, before being business wargames, wargames were developed to propose visual solutions to recreate a specific war situation. Now, wargames for civilians exist, with a particular kind: board wargames, which we found very interesting for information display. In this paper, we develop a methodology to apply a board wargame tool for a market situation. This methodology contributes to creative competitive intelligence (or creative watch) a new kind of competitive intelligence, in the sense it participates to information discovery that directly contributes to the creation and innovation process.