When Is a Tree a Hedge?

Linear models dominate quantitative thinking in investments. The weighted sum of a relevant set of factors, typically determined from a regression or discriminant analysis, determines model expectation. This approach is quite different from the heuristic style of decision making people tend to prefe...

Full description

Saved in:
Bibliographic Details
Published in:Financial analysts journal 1994-11, Vol.50 (6), p.75-81
Main Author: Mezrich, Joseph J.
Format: Article
Language:English
Subjects:
Carts
Classification
Decision
Decision trees
Dependent variables
Economic theory
Financial advisers
Financial investments
Forecasting models
Information classification
Investment decision
Investment policy
Investment strategies
Linear models
Linear regression
Options trading
Price volatility
Put & call options
Regression analysis
Volatility
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Linear models dominate quantitative thinking in investments. The weighted sum of a relevant set of factors, typically determined from a regression or discriminant analysis, determines model expectation. This approach is quite different from the heuristic style of decision making people tend to prefer. For example, when a patient is admitted to a trauma center, the attending physician determines the course of treatment for the emergency by going through a set of hierarchical decisions. Some things are more important than others, and they get considered first. Is there a pulse? is clearly a question to be asked first. Then follow questions of decreasing significance in deciding the course of treatment for the patient. The decision model for the physician is a decision tree, with an appropriate set of "if-then" rules. The decision tree is determined in a heuristic fashion based on experience and is preferred to a quantitative, weighted-sum approach, as in regression or discriminant analysis. The best investment decision making is more along the lines of using the right decision tree than using the right linear model.
ISSN:0015-198X
1938-3312
DOI:10.2469/faj.v50.n6.75