Neural Network Stock Pricing Model

NightScope is an experimental stock valuation model based on Artificial Neural Network (ANN) technology. Using data on almost 3500 U.S. Stocks, the ANN models the relationship between a core set of fundamentals and current price. Premiums or discounts to the model's price are reported and assumed to contain the beliefs, rational and irrational, of investors. The difference between market prices and fundamental values, as determined by the NightScope model, can be directly observed. Value investors rely on finding stocks trading away from fundamental value. Now, there is a powerful new way to find fundamental values -- NightScope: a nonlinear, data driven, stock pricing model.

Currently, the model is not maintained on a regular basis. I believe the underlying ideas are valid and contain useful advances in the application of ANN technology to common stock valuation; however, I don't have time to maintain a current stream of updates.

• How does it work?: Background on the pricing methodology.
• Sources and Ideas: References used to develop the site.
• June 30, 1996 Price Model: Pre-July correction prices.
• July 31, 1996 Price Model: July correction aftermath and commentary.
• July 31, 1996 Price Model - with August 21 prices: July 31 Model with updated prices for calculating premiums and discounts (with commentary).
• September 30, 1996 Price Model: At long last, an update on the model for recent fundamental data and prices.
• Zytek: A great source for internet services and consultation.

For more information, or to send comments, contact me. To visit my homepage, click here.

Disclaimer

Use this information at your own risk.