An application of DATALOGIC/R knowledge discovery tool to identify strong predictive rules in stock market data

An application of a methodology for discovering strong probabilistic rules in data is presented. The methodology is based on an extended model of rough sets called variable precision rough sets model incorporated in DATALOGIC/R knowledge discovery tool from Reduct Systems Inc. It has been applied to analyze monthly stock market data collected over a ten year period. The objective of the analysis was to identify dominant relationships among fluctuations of market indicators and stock prices. For the purpose of comparison, both precise and imprecise, strong and weak rules were discovered and evaluated by a domain expert, a stock broker. The evaluation revealed that the strong rules (supported by many cases) essentially confirm the expert's experiences whereas weak rules are often difficult to interpret. This suggests the use of rule strength as the primary criteria for the selection of potentially useful predictive rules.