Research On Chance Discovery Algorithm And Its Application

Chance Discovery is a new research area. It is important to the research content of Chance Discovery, which is how to discover events or scenarios these are important to decision agent but rare or hidden from data set.By the analysis and study of the existing chance discovery algorithms, advantages and disadvantages of the classical algorithms are analysed and new ideas and algorithms are proposed in this paper. The formal description of chance discovery and the system architecture are also given in this paper.The main researches of this paper are on the following:(1) Research on the formal description of chance discovery. The modal logic is introduced and extended, and the logical language LDKa based on modal logic is given. The chance discovery is described by using the LDKa language alse. In order to get the better description of the concept and characteristics of chance and chance discovery, decidability and satisfiability proof procedure of the logical language LDKa is given, and the validity of LDKa is clarified further.(2) Research on discovery algorithm of rare events. The main content of this research is how to discover the rare and significant events from the given data set. Optimize algorithm by Small World Net theory; it effectively enhances the accuracy and performance of algorithm.(3) Research on discovery algorithm for hidden events. The main content of this research is how to discover the hidden and significant events from the given data set. These hidden events can give the better way to understand and interpret existing phenomena. By introducing Heat-Annealing theory, it increases the scope of algorithm application, at the same time, increases the proportion of explainable hidden events.(4) Research on the software architecture. The Intelligent Agent technology is introduced into chance discovery architecture. While retaining the advantages of double helix model, the system implementation plan is provided.In this paper, patent specifications, as the application background, their scenario graph model is established. Based on this scenario graph model, chance which is rare or hidden in the data set is discovered. This chance can be regard as the opportunity for pioneering a new cause, and the unique advantages of chance discovery algotithms and architecture are clarified.