Title: Neural networks and technical analysis for price prediction: the case of Borsa Italiana SP MIB
Authors: Angelo Corelli
Addresses: Center of Excellence for Research in Finance and Accounting (CERFA), American University in Dubai, United Arab Emirates; Finance and Accounting Department, School of Business and Administration, American University in Dubai, United Arab Emirates
Abstract: The paper analyses the relationship between common technical analysis indicators and the returns of an index for the period considered. It is expected to find correlation between indicators and index prices, as well as showing clear patterns and potential strategies for investment analysis and portfolio management. As an innovative methodology a mixed analysis is carried out, trying to combine classic signals offered by the indicators with the power of neural networks. The neural network plays an important role in that allows for an accurate regression with efficient error minimisation, while giving indications about the concentration of results obtained around some reference values. Through a simple hidden-layer, back-propagation algorithm, regressions give interesting result, in term of the forecasting potential of the analysed indicators. The final step of the project is to conclude about results and summarise the indication coming from the multivariate stage analysis, commenting on the power of the indicators to reveal potential investment opportunities.
Keywords: technical indicator; neural network; backpropagation; steepest descent; normalisation.
International Journal of Business Continuity and Risk Management, 2020 Vol.10 No.1, pp.1 - 22
Received: 13 Sep 2018
Accepted: 20 Mar 2019
Published online: 04 Mar 2020 *