V.2:2 (55-59): Optimizing the Maximum Entropy Method by Anthony W. Warren, Ph.D.
Product Description
Optimizing the Maximum Entropy Method
by Anthony W. Warren, PhD
This article is the second in a series on the Maximum Entropy Method (MEM) and its use in technical analysis for short data length spectrum estimation and short term forecasting. In the first article (January 1984 issue of T.A.) we introduced the main ideas behind MEM and illustrated the relationship between the MEM order or number of days used in MEM forecasts, and the resolution in the spectrum. In the following article we discuss the critical process of optimizing the MEM fit coefficients to obtain reliable models for spectrum estimation. This process is crucial for obtaining high resolution spectra from short data sets (say 30 to 80 samples of price history data), and for obtaining forecasts consistent with recent data trends and cycle lengths. The optimizing methods discussed here represent the product of both academic research on the Maximum Entropy Method and the author's research on application of this method to stock and commodity data. The result of this effort is a semi-automated optimization algorithm (to be published in a subsequent issue of T.A.) which is relatively easy to use, and which in most cases will produce reliable spectra and forecasts. The application of this method to trend channel forecasting and usage within trading systems will be discussed in a subsequent article.
MEM OPTIMIZATION PARAMETERS
There are three parameters which may be selected independently in the author's optimizing method:
* A prefilter-and-sample parameter- D
* The MEM prediction order- M
* The data length- N
This article is the second in a series on the Maximum Entropy Method (MEM) and its use in technical analysis for short data length spectrum estimation and short term forecasting. In the first article (January 1984 issue of T.A.) we introduced the main ideas behind MEM and illustrated the relationship between the MEM order or number of days used in MEM forecasts, and the resolution in the spectrum. In the following article we discuss the critical process of optimizing the MEM fit coefficients to obtain reliable models for spectrum estimation. This process is crucial for obtaining high resolution spectra from short data sets (say 30 to 80 samples of price history data), and for obtaining forecasts consistent with recent data trends and cycle lengths. The optimizing methods discussed here represent the product of both academic research on the Maximum Entropy Method and the author's research on application of this method to stock and commodity data. The result of this effort is a semi-automated optimization algorithm (to be published in a subsequent issue of T.A.) which is relatively easy to use, and which in most cases will produce reliable spectra and forecasts. The application of this method to trend channel forecasting and usage within trading systems will be discussed in a subsequent article.
MEM OPTIMIZATION PARAMETERS
There are three parameters which may be selected independently in the author's optimizing method:
* A prefilter-and-sample parameter- D
* The MEM prediction order- M
* The data length- N
FOR THOSE ORDERING ARTICLES SEPARATELY:
*Note: $2.95-$5.95 Articles are in PDF format only. No hard copy of the article(s) will be delivered. During checkout, click the "Download Now" button to immediately receive your article(s) purchase. STOCKS & COMMODITIES magazine is delivered via mail. After paying for your subscription at store.traders.com users can view the S&C Digital Edition in the subscriber's section on Traders.com.
| Take Control of Your Trading. |
| Professional Traders' Starter Kit |
| All these items shown below only $299.99! |
| Click Here to Order |