Computing Environments for Data Analysis. Part 1. Introduction
Abstract
Statistics has long been thought of as applied mathematics. Certain parts of it, especially data analysis, could easily be viewed as applied computation. In this context, different research issues assume importance, in particular, the design and implementation of computing environments for data analysis. Analyzing data requires computing, whether with paper-and-pencil or with a Cray super computer. The computing environment determines what sorts of statistical methods are practical. More importantly, the statistician's unconscious assumptions, or mental model of the computing environment determines the kinds of new statistical methods that are likely to be invented. Most current research in statistical computing is based on a batch processing model of computing environments that was appropriate twenty years ago. With a few exceptions, statisticians have not addressed the implications current and future developments in scientific computing environments. This paper is the first in a series that discuss local networks of graphics workstations as environments for statistical computing. In this, first part, we provide general background and motivation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1984
- Accession Number
- ADA148677
Entities
People
- Jan Pedersen
- John A. Mcdonald
Organizations
- Stanford University