The BIG Picture
- A unified perspective,
showing how and why Jurik's modules work well as building blocks for reliable
low-lag indicators. Includes graphics. Author: Mark Jurik
Why Use JMA ?
- Outlines the four
basic benchmarks for judging the quality of moving averages with regard to
financial trading. Compares JMA to classical and modern filter designs.
Includes graphics. Author: Mark Jurik
Evolution of Moving
Averages
- Summarizes the recent
evolution of moving average filter design. Compares popular versions to a set
of ideal performance features. Regarding how well filters process noisy time
series data with price-gaps, report shows that the latest designs are getting
very close to theoretical performance limits. Includes graphics. Author: Mark
Jurik
Relating Neural Networks
to Statistical Methods
- Summarizes the
relationship between neural nets and modern statistical methods. No
mathematics. Author's conclusion is that "most neural networks that can
learn to generalize effectively from noisy data are similar or identical to
statistical methods." Also lists neural net models that have no close
relatives in the existing statistical literature. Appended to this document is
a comparison between verbal jargon used by neural netters and statisticians.
Author: Warren Sarle
Neural Networks for
Trading the Markets: Primer
- A brief introduction
to the use of neural networks suitable for futures forecasting. Author: Don W.
Fitzpatrick
Neural Networks for
Trading the Markets: Case Study #1
TITLE: Neural Nets for Personal
Investing
- This version,
submitted to us by the author, is an adaptation of his original article
submitted to HEURISTICS: The Journal of Intelligent Technologies, to be
published in their special issue: Neural Networks for Financial Systems, v9,
#1. Reviews the development and results of a neural-net based trading system.
Author: William Arnold
Neural
Networks for Trading the Markets: Case Study #2
TITLE: Financial Time Series Forecasting
by Neural Networks
- Compares two different
neural network training algorithms used to model the time series of companies
on the Shanghai Stock Exchange. Shows that the Conjugate Gradient Descent
algorithm is better than classic Gradient Descent. Authors: CHAN Man-Chung,
WONG Chi-Cheong, LAM Chi-Chung -- (Hong Kong Polytechnic University)
Overview of
BackPercolation
- A non-mathematical
overview of the philosophy behind the design of the BackPercolation method of
training Perceptron-based neural nets. Author: Mark Jurik
Some Programming
Issues in TradeStation EasyLanguage
- This document
illustrates how TradeStation may produce counter-intuitive results when calling
Easy Language functions. Alternative code that avoids the problem is provided
and each case clearly explains why one method works and the other does not.
Lastly, examples are provided showing how to avoid these situations when using
studies from Jurik Research. -- Author: Mark Jurik
Series/Simple Functions
in Easy Language
- Explains the
fundamental difference between two types of Easy Language functions in
TradeStation. Charts included. Author: Mark Jurik
Optimal Forecast
Horizon
- Leading indicators
require data with low noise and low lag, because that combination yields the
widest window of time in which a forecast can be accurate. This paper briefly
touches on chaos theory to present the notion of any time series having an
"optimal" forecast horizon. Author: Mark Jurik
Classification Tree
of Modeling Techniques
- This one page diagram
shows all the modeling methods arranged in a hierarchical tree, where results
from one method feed into other methods. Great for getting the big picture on
modeling methods and how they relate. Author: Unknown
Neural Networks: Myths and
Reality (web link)
- So what is neural
network technology, what should and what shouldn't a trader expect from it if
he selects to use it to achieve his trading goals?
|
To download each PDF file listed below, right click
on hyperlink and use menu. 
