Abstract
The financial market has always been an interesting target for researcher, the prediction of stock market is the most attractive topic. Recent research tries to involve the textual data such as news and blog into financial area. In this paper, we provide a method to automat- ically extract the representative concepts from amount of news data, and use stock index data to verify the influence of each concept. Furthermore, we also provide few methods by using events to predict the future trend. The distinguishing feature of our paper is to use an event angle to explain the trend of stock market, which is not common in the area. The technology is used in this paper include the statistics, Genetic Algorithm, K-nearest neighbor, frequent itemset and regression model. We also invent a pheromone prediction model for predicting the stock prices.