Some Topics in Data Analysis and Stochastic Modeling

Abstract

The advent of high speed computing on the PC provides new possibilities for data based modeling. One key is simulation. Simulation gives us a venue for dealing with parameter estimation in stochastic models not previously tractable. We consider examples from oncology, economics, statistical process control and epidemiology. We essentially consider two realities, the first consisting of random nonparametric interpolations of the data and the second random implementations based on a mathematical model. We use simulation to change the model parameters to bring the reality of the model to consistency with that of the data. Also, high speed computing enables us to carry out nonparametric data analysis in higher dimensions. The key here is to build algebraic rather than geometrical algorithms. We show how nonparametric data analysis in high dimensions (say greater than three) should generally not be treated as one of nonparametric density estimation but rather should start with finding centers of relatively high density and then use parametric approximations in the neighborhoods of these modes.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1999
Accession Number
ADA370056

Entities

People

  • James R. Thompson

Organizations

  • Rice University

Tags

Communities of Interest

  • Ground and Sea Platforms

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Consistency
  • Data Analysis
  • Data Science
  • Databases
  • Economics
  • High Density
  • Information Science
  • Mathematical Models
  • Models
  • Natural Gas
  • Oncology
  • Simulations
  • Statistical Processes
  • Statistics
  • United States

Fields of Study

  • Mathematics

Readers

  • Computational Modeling and Simulation
  • Educational Psychology
  • Statistical inference.