Causal Links Analysis Based Conceptual Model for Stocks

In recent years, people from academic and practice fields put more and more effort to study various stock phenomena using techniques from AI and data mining. However, the way they process data generated from phenomena has two problems, the first is that people just focus on relations among observable data, not relations among entities which are more stable than the former ones. The other problem is that people now usually store data in multi-relational tables in databases, if there are no links knowledge provided for these tables, it is hard for them to draw a compound model for their studying certain phenomenon. Thus in this paper, a conceptual model is proposed to manage the links knowledge for stock domain in a formal way. It include an ontology of classed of links and inductive rules complementing to ontology because of the representation limitation of ontology. With this conceptual model, system development and data mining for certain stock phenomena can be supported by more stable background knowledge, particular the links information among entities in stock domain.