NXCS: Hybrid Approach to Stock Indexes Forecasting

摘要: In this chapter, a hybrid approach for stock market forecasting is presented. It allows to develop mixture of experts, each consisting genetic classifier and an associated artificial neural network. The resulting experts have been applied using technical trading rules as inputs other inputs—in particular past quotations—for the networks. particular, former are used find quasi-stationary regimes within financial data series, whereas latter assigned task making “context-dependent” predictions on next day trend market. To end, novel kind feedforward network has defined, allowing implement suitable predictors without being compelled exploit more complex architectures. Test runs performed some well-known indexes, also taking into account commissions. tests pointed good capability proposed approach, which repeatedly outperformed buy-and-hold strategy.