Mining Event Associations using Structured Data and Classifiers

Event is a widely used concept these years. Many areas such as Natural Language Process, Information Retrieval have used event as the basic information unit in their research. So, the mining of event association is very necessary for our research. And it plays an important role business intelligence and researches of relations between events. Usually events are associated with others when they often occur in the vicinity of others or co-occur in the same context. However, there are some implicit associations we cannot mine only from sequence or context. In this paper, we aim to find associations of events under the background of Data Integration Systems. By using the structured information of data integration system, the background information of entities can be extracted to classify events. So we classify the events into different categories which makes it possible to mine the statistical information from event sequence. Furthermore, we generalize the association between event entities to predict the implicit association in our algorithm. We validate our method with experiments and results show the useful information in the area of business intelligence.