US Corporations are required by the Securities and Exchange Commission to file annual and interim disclosures on the Electronic

Data Gathering, Analysis and Retrieval (EDGAR) system. Shareholders and investors attempt to analyse these disclosures in the hope

of predicting the likely share price response. We are investigating the correlation between the release of disclosure information and the

share price response and present three experiments in which a Decision Tree System and an Inverted Index System are trained against

a collection of disclosures. The disclosures have been tokenised and contain only the most frequently occurring compounds. The

Decision Tree System equalled or outperformed the Inverted Index System in two of the three experiments.