Review of ‘4Thought’

Review of ‘4Thought’

0.00 Avg rating0 Votes
Article ID: iaor19942361
Country: Netherlands
Volume: 10
Issue: 1
Start Page Number: 35
End Page Number: 41
Publication Date: Mar 1994
Journal: International Journal of Forecasting
Authors: ,
Keywords: time series & forecasting methods, artificial intelligence
Abstract:

The use of neural nets for forcasting and modelling has aroused considerable interest in recent years. 4Thought is a computer package which is aimed at the business user and can be used on a PC, preferably a fast 386 or 486 machine with a co-processor. 4Thought is menu driven and operates in a windows environment. It comes with a manual and contains a number of worked examples. However, before discussing the program, the authors first describe the neural net models which it uses and explain how these are related to standard statistical procedures. (In common with most of the neural net literature, the 4Thought manual makes no attempt to relate what is being done to standard techniques.) Having described the program, the authors go on to discuss their experience with modelling and forecasting some real time series. This makes it possible to get some feel of how the neural net techniques used in 4Thought compare with other methods. The review therefore examines two issues. First, how useful is 4Thought as a program for modelling neural nets? Second, and this is clearly a much wider issue, how effective are neural nets (at least as implemented in 4Thought) for modelling economic and financial time series?

Reviews

Required fields are marked *. Your email address will not be published.