Mining Technical Patterns in The U. S. Stock Market through Soft Computing

Abstract

Technical analysis has been a part of financial practice for many decades. One of the most challenging areas in technical analysis is the automatic detection of technical patterns that are similar in the eyes of expert investors. In this chapter, we propose a soft computing based approach for technical analysis. By introducing the inter ‘ intra fuzzification into an automatic pattern detection and analysis process, we incorporate human cognitive uncertainty into the technical analysis domain. The importance of fuzzy technical patterns on investment decisions is confirmed through a neural network based saliency analysis. Using a random sample of U. S. stocks, we find that our approach is able to detect subtle differences within a clearly defined pattern. Our results suggest that such subtle differences could be a source of controversy surrounding technical analysis. Compared with existing visual technical pattern analysis approaches, our soft computing based approach offers superior precision in detecting and interpreting technical patterns.