2005 | OriginalPaper | Chapter
Empirical Approach as a Scientific Framework for Data Analysis
Author : Shizuhiko Nishisato
Published in: Data Analysis and Decision Support
Publisher: Springer Berlin Heidelberg
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
The traditional statistical procedure is typically based on the notion that data are a random sample from the normal population. Knowing that the bivariate normal distribution has only three parameters, mean, variance and linear correlation, the use of the normal distribution as an analytical framework leads to what we call linear analysis. This paper starts with discarding the normal distribution assumption, and then advocates total reliance on data in hand. In the social sciences, we face the majority of data to be categorical. To explain the data exhaustively, it is almost necessary to employ an empirical approach without any prior assumptions such as the distribution, levels of measurement or models for data. This framework opens up the possibility of capturing nonlinear information and multidimensionality in data, as well as of cautionary wisdom for the popular data reduction approach to data analysis.