Introduction to Neural Networks for C#, Session 11

Session Material

In this class session we use predictive neural networks to predict the S&P 500. To attempt to predict the S&P 500 we use a window size of 10. However, we do not just use past data from the S&P, we also use the current prime interest rate. This causes us to have 20 input neurons. 10 neurons from the last 10 periods of the S&P. Also 10 more neurons from the last 10 periods of the prime interest rate. We have a single output neuron that attempts to predict the next value of the SP&P. For more information on how the network is constructed, refer to the book or the videos.

Programming Assignment 2

For programming assignment 2 you should expand upon the mid-term program by adding incremental pruning to determine the optimal number of hidden layers and hidden neurons. You should loop up through the neuron counts for hidden layer 1 and hidden layer 2. Do not do a full training cycle, it will take too long. Rather just train a set number of iterations. Determine which configuration trains the best in the number of iterations you chose.

Videos for this Session

Video	Title
	Introduction to Neural Networks for C#(Class 11/16, Part 1/5)
	Introduction to Neural Networks for C#(Class 11/16, Part 2/5)
	Introduction to Neural Networks for C#(Class 11/16, Part 3/5)
	Introduction to Neural Networks for C#(Class 11/16, Part 4/5)
	Introduction to Neural Networks for C#(Class 11/16, Part 5/5)