Realization of a news dissemination agent based on weighted association rules and text mining techniques

In this work, a financial news headline agent is proposed to assisting the investors in deciding to buy and to sell stocks in Taiwan market after receiving the essential real-time news headline disseminated by the agent. Weighted association rules and text mining techniques are used to derive the significance degree of each newly arrived news headline on the fluctuation of Taiwan Stock Exchange Financial Price Index on the next trading day. The experimental results reveal that the proposed work indeed achieves significant performance and demonstrate its feasibility in the applications of real-time information dissemination, such as financial news headlines via Internet.