A Fuzzy Based Recommendation System for Stock Trading

Predicting prices, as well as giving buy and sells recommendations in a Stock market, is considered difficult given the very complex behavior of the price itself. It is considered dynamic, non-linear and stochastic. In similarity to other fields, stock trading firms may profit from using computational intelligence systems, reducing the analysis time and enhancing the accuracy of recommendations. This paper shows a Fuzzy trading system based in technical analysis developed in two steps: primarily, a fuzzy trading system based on common technical indicators used by technical analysts is proposed, and after, the system is supported with a price prediction methodology. The first result shows that not all the stock tickers are eligible for a simple indicator operation, which may result in a severe loss. The usage of a price prediction methodology shows real improvement on the recommendation system.