A novel approach for candlestick technical analysis using a combination of the support vector machine and particle swarm optimization

Purpose In this research, the main purpose is to use a suitable structure predict trading signals of stock market with high accuracy. For purpose, two models for analysis technical adaptation were used in study. Design/methodology/approach It can be seen that support vector machine (SVM) particle swarm optimization (PSO) where PSO as fast and accurate classification search problem-solving space finally results are compared neural network performance. Findings Based on result, authors say both new trustworthy 6 days, however, SVM-PSO better than basic research. The hit rate 77.5%, but networks (basic research) 74.2. Originality/value approaches (raw-based signal-based) have been developed generate input data model: raw-based signal-based. comparison, considered percentage correct predictions 16 days.