Regression in Analysis
Abstract
In regression analysis, the goal is to determine the values of parameters for a function to best fit a set of data observations. Put another way, regression attempts to best describe what inputs result in a given output. Though there are many complex forms of regression models, the simplest is a linear regression model. In this paper, the author uses linear regression analysis to determine what factors may or may not contribute to the emplacement of an improvised explosive device (IED) in and around Baghdad. The regression model he uses takes into account the following variables: the socioeconomic status of an area, previous IED activity there, time, and unattributed factors. He reminds the reader that regression analysis is nothing more than another tool available to the analyst. It is not, and should not be, the analysis itself. Also, a regression model is a very specific thing. While the intent is to create a model and analyze a data set so as to better predict, one must realize that, ultimately, the model only tells the analyst about that specific data set. It is up to the analyst, and those creating the models, to determine whether or not effective predictions can be made. Lastly, regardless of whether one understands regression a little bit or not at all, it is important to remember that the whole point is to make improvements. While the IED model presented here is clearly immature, incomplete, and overly simple, it is a beginning. "Good enough" is a phrase that should never enter the analyst's lexicon. To do so is to put soldiers at risk willingly. Improvement, not perfection, is the goal, and the addition of regression analysis to the analyst's tool kit would certainly be a vast improvement.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 12, 2008
- Accession Number
- ADA494895
Entities
People
- Kevin Burke