Material Detail

Modelling Financial Time Series using Grammatical Evolution

Modelling Financial Time Series using Grammatical Evolution

This video was recorded at International Workshop on Advances in Machine Learning for Computational Finance (AMLCF), London 2009. The traditional models of price, and its statistical signatures are often based on limiting assumptions, such as linearity. Moreover, the model developer is faced with the model selection problem, and model uncertainty. In this paper we introduce a method based on Grammatical Evolution (GE) to evolve models for predicting financial returns, and we examine the profitability of these models. Our empirical analysis demonstrates that for some securities our method is able to produce models of return that are lead to more profitable trading compared with an Autoregressive model picked using Aikake Information Criterion (AIC), under the assumption of frictionless markets.

Quality

  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collection (1) Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material

Comments

Log in to participate in the discussions or sign up if you are not already a MERLOT member.