Short-term prediction of the Romanian stock market benchmark index using genetic programming

The evolution of stock market indices is one of the most intensely studied subjects, given that in this particular field successful insights can offer both academic and financial benefits. Understandably, most focus lies on developed markets that drive economies, with less interest in frontier or emerging markets such as the Romanian stock market. In this paper, we offer an analysis of the connections between the Romanian stock market index BET and major indices from the USA (NASDAQ, S&P500, DIJA), Western Europe (FTSE100, CAC40, DAX) and Central and Eastern Europe (ATX,BUX, WIG20). The analysis explores the use of genetic programming models for the prediction of index values using short-term intervals and compares them with standard regression models. We find that most methods perform in an almost similar manner and that the best numerical results are obtained when considering the Central and Eastern European indices.