Collective Artificial Intelligence for Mechanical Technical Analysis
Because technical analysis uses only historical price data, it has affinity with machine learning approach and is often called modern technical analysis or mechanical technical analysis. In general, machine learning method is categorized as an artificial intelligence (AI), which is really a hot topic in the world since Google’s Al (AlphaGo) beat the world Go champion. Then, it has been called FinTech to apply Al for financial business intelligence.
Professor Suzuki introduces some ideas to develop Al algorithms for technical analysis, such as collective intelligence and abnormally detection. The collective intelligence can enhance the predictive power by integrating many neural networks and can select the most reliable stock by following their consensus. Then, the autoencoder is a new neural network used for the deep neural network, and he applies it to the abnormally detection of noisy financial markets. To confirm the validity of these ideas, Professor Suzuki performs investment simulations and statistical significance tests with real stock price data.
![KNILDMpecKg.jpg]