Close
ANNFS-P06
(PGP 2006-08 : Term-V)

Artificial Neural Networks: Applications to Finance and Strategy (ANNFS)

Course Outline, PGP II, Term V



Course Instructor
K. M. Rajesh (rajesh@ximb.ac.in)

Course Objective
· To introduce ANN as a data modeling tool
· To familiarize the participants with different neural network learning methodologies
· To acclimatize the participants with some modeling problems in finance using neural networks and traditional methodologies
· To gain a good understanding of at least one of the above problems through a course project
· To overcome the mind block of looking at neural networks as black boxes

Course Contents
Motivation to use ANN.
Introduction to ANN architectures, learning, and to some typical problems where the technique can be applied.
Modeling issues - problem representation, variable selection, preprocessing of data, architecture construction, control parameters, cross-validation, performance metrics.
Fundamental building blocks of ANN
Learning algorithms in ANN - supervised, unsupervised, self-organization, competitive;
Neural network architectures - feed-forward networks and back-propagation; radial basis functions; constructive learning algorithms; adaptive resonance theory; associative memory networks.
Applications - Bankruptcy prediction, mergers/acquisitions prediction, credit rating, derivatives pricing, stock returns prediction, modeling stock indices, currency exchange rate prediction, portfolio management, testing market efficiency and coherence.

Pedagogy
The class lectures would discuss ANN theory.
Article presentations, assignments and project would be done in teams of 2 (or 3) students. They will be oriented towards application of neural network in financial problems.

Evaluation
Mid term (15%): Based on class lectures
Presentations (10%): Based on articles distributed for reading
Assignments (15%): Simple problems in neural network modeling
Project (25%): Detailed exposition of problem solving using ANN of a practical modeling problem in finance.
End term (35%) : Based on the entire course coverage.

Project will have three phases of submission. The first phase would be to identify a problem, team members, and data sources. The second phase would involve studying various possible neural network and non-neural network methodologies to solve the problem. The third phase would involve actually trying to solving the problem using the methodologies mentioned in the second phase and analyzing the results. More details about the project will be discussed in the first class.


References
1. Mehrotra, Mohan and Ranka, Elements of artificial neural networks, MIT press.
2. Hassoun, Fundamentals of artificial neural networks, Prentice Hall of India.
3. Haykin, Neural Networks : A Comprehensive Foundation, 2/e, Pearson Education.
4. Apostolos-Refenes, Neural networks in the capital markets (Ed), John Wiley.
5. Azoff, Neural networks in time series forecasting of financial markets, John Wiley.
6. Trippi and Turban, Neural networks in finance and investing, Heinemann Asia.
7. Focardi, Modelling the market, Frank Fabozzi & Associates.
8. Zenios, Financial optimization, Cambridge University Press.
Created By: Lingaraj Pattanaik on 06/05/2006 at 09:09 AM
Category: PGP-II Doctype: Document

...........................