- author = "Michael F. Korns",
- title = "Trading Volatility Using Highly Accurate Symbolic Regression",
- booktitle = "Handbook of Genetic Programming Applications",
- publisher = "Springer",
- year = "2015",
- editor = "Amir H. Gandomi and Amir H. Alavi and Conor Ryan",
- pages = "531--547",
- keywords = "genetic algorithms, genetic programming, Value investing, Symbolic regression, Logistic regression, Nonlinear regression",
- isbn13 = "978-3-319-20882-4",
-
DOI = "
doi:10.1007/978-3-319-20883-1_21",
-
abstract = "Research efforts, directed at increasing the accuracy
and dependability of Symbolic Regression (SR), have
resulted in significant improvements in symbolic
regression's range, accuracy, and dependability.
Previous research has also demonstrated the
practicability of estimating corporate forward 12 month
earnings, using advanced symbolic regression. In this
paper we put these prior results and techniques
together to select a 100 stock semi-passive index
portfolio, extracted from the Value Line Timeliness
stocks (Value Line), which delivers consistent
performance in both bull and bear decades and we will
compare its performance to the Standard & Poors 100
index.
We intend to produce our 100 stock semi-passive index buy list on a weekly basis using automated forward 12 month EPS (ftmEPS) prediction involving the analysis of many securities, involving multiple training regressions each on hundreds of thousands of training examples. Plus the timeliness issue will require that our analytic tools be strong and thoroughly matured. The 100 stock buy list will be the foundation for a new semi-passive Value Line 100 index fund which should have great appeal to many high net worth clients, enjoy low management costs, and be easily acceptable to the compliance and regulatory authorities.
Valuation of Value Line securities via their forward 12 month price earnings ratio (ftmPE) is a very common securities valuation method in the industry. Obviously the ftmPE valuation depends heavily on the estimate of forward 12 month corporate earnings per share (ftmEPS). Several obvious inputs to the ftmEPS prediction process are the past earnings time series plus one or more analyst predictions.
Valuation via ftmEPS is a necessary but not a sufficient attraction for a semi-passive index fund. So we will introduce the advantages of trading volatility. Our thesis will be that emotional trading patterns tend to make markets less efficient.
The efficient market hypothesis depends upon equal access to information and rational trading patterns. Trading on insider information is illegal in most developed securities markets; however, trading when others are emotional is unregulated. In this paper we will develop a set of factors, all of which incorporate a measure of volatility indicating possible overly emotional trading patterns. The theme of our new semi-passive index fund will be Buy value from those who are selling in a highly emotional state.",
