Data stream is a type of dynamic data, which is driven by some hidden contexts and may change with time. Generally speaking, the change implies some event in the real world. To detect change over data stream timely and accurately has a quite wide range of practical applications, and has been one of the hot topics in data stream mining. In this paper, a comprehensive survey on data stream change detection was presented, and the key task of change detection was introduced. An inductive unifying framework of change detection process was also given. Besides, a variety of already existing change detection approaches and their features were reviewed and evaluated in detail. Finally, the research outlook of data stream change detection was discussed.

Based on the characteristics of high-speed maglev train line structure and operation control system, this paper set up a model for the high-speed maglev train diagram of double-track lines, and put forward the knitting algorithm to resolve the model. The algorithm drew the up and down maglev train running lines in chronological order of trains’ departure time from stations, which tended to do some knitting. When meeting the conflict, the algorithm resolved it by changing the station tracks or increasing the departure time intervals, at last it gained global optimization maglev train diagram step by step. A numerical example shows that the method can draw the maglev train diagram quickly and effectively.

In order to solve the semantic conflict of emergency cases understanding in Case-Based Reasoning (CBR), an emergency case ontology model based on CBR was defined on the basis of the extended ABC ontology model according to the analysis of emergency cases, and the key elements of the model were described in detail, such as concepts, relations, axioms and instances. Finally, the model was validated by the analysis on the case of the nuclear crises of the first nuclear power station in Fukushima.