Trading ebook Finance - Neural Prediction of Weekly Stock Market Index

425 043 60 40 20 delay 53 0 20 feature extraction

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: .41 -0.415 -0.42 -0.425 -0.43 -60 -40 -20 Delay-53 0 20 Feature extraction Given, the above analysis the inputs to the neural network are: • The stock index value for the present week (all values are weekly closing values) • the weekly long term interest rate, delayed by 25 weeks, • the weekly short term interest rate, delayed by 20 weeks and • the weekly short term interest rate, delayed by 48 weeks. 7 The stock market future value also depends on the trend of the market as well as the trend of the interest rates. To include this information, additional inputs are used. These are: • The difference between the present week’s index and that of the previous week (referred to as the first difference). Since, this difference can be positive or negative, it is encoded (as done in [7]) using two inputs. The first one is the absolute value of this first difference and the second one is arbitrarily chosen as: 0.8 if the difference is positive, 0.2 if the difference is negative and 0 if there is no change. This encoded sign of the difference is h...
View Full Document

This note was uploaded on 07/20/2012 for the course ECON 203 taught by Professor Girishdev during the Spring '12 term at Indian Institute of Technology, Kharagpur.

Ask a homework question - tutors are online