A practical investment problem is defined as follows: decide today whether to switch in or out of an S&P 500 index fund, where an initial $10,000 investment will be held for 13 weeks before making the next decision. At the end of the 13-week holding period, the process is repeated. The alternative investment is risk-free and yields a generally available compound annual rate of return of 5%. The performance of 3 strategies are compared: (1) buy-and-hold, (2) dollar-cost averaging, and (3) learning network advisors. The decisions made by the learning systems are based on at most nine inputs: the S&P 500 13-week holding-period yield at the close of today, and 8 past 13-week yields spaced 13 weeks apart. It is shown that very simple wavelet and polynomial neural networks are able to match or exceed the limit of performance implied by the efficient market hypothesis as represented by the buy-and-hold strategy.
[1]
Ingrid Daubechies,et al.
The wavelet transform, time-frequency localization and signal analysis
,
1990,
IEEE Trans. Inf. Theory.
[2]
Stephen F. Witt,et al.
Portfolio theory and investment management
,
1994
.
[3]
Roger L. Barron,et al.
Applications of Polynomial Neural Networks to FDIE and Reconfigurable Flight Control
,
1998,
Proceedings of the IEEE 1998 National Aerospace and Electronics Conference. NAECON 1998. Celebrating 50 Years (Cat. No.98CH36185).
[4]
Mohamad T. Musavi,et al.
On the training of radial basis function classifiers
,
1992,
Neural Networks.
[5]
John L. Casti,et al.
Searching for Certainty: What Scientists Can Know about the Future
,
1992
.