Classification and Powerlaws: The Logarithmic Transformation - Computer Science > Information RetrievalReport as inadecuate




Classification and Powerlaws: The Logarithmic Transformation - Computer Science > Information Retrieval - Download this document for free, or read online. Document in PDF available to download.

Abstract: Logarithmic transformation of the data has been recommended by the literaturein the case of highly skewed distributions such as those commonly found ininformation science. The purpose of the transformation is to make the dataconform to the lognormal law of error for inferential purposes. How does thistransformation affect the analysis? We factor analyze and visualize thecitation environment of the Journal of the American Chemical Society JACSbefore and after a logarithmic transformation. The transformation stronglyreduces the variance necessary for classificatory purposes and therefore iscounterproductive to the purposes of the descriptive statistics. We recommendagainst the logarithmic transformation when sets cannot be definedunambiguously. The intellectual organization of the sciences is reflected inthe curvilinear parts of the citation distributions, while negative powerlawsfit excellently to the tails of the distributions.



Author: Loet Leydesdorff, Stephen Bensman

Source: https://arxiv.org/







Related documents