“Left shoulder” detection in Korea composite stock price index using an auto-associative neural network

Abstract

We propose a neural network based “left shoulder” detector. The auto-associative neural network was trained with the “left shoulder” patterns obtained from the Korea Composite Stock Price Index, and then tested out-of-sample with a reasonably good result. A hypothetical investment strategy based on the detector achieved a return of 124% in comparison with 39% return from a buy and hold strategy.