The Research Of OLAM Based On Time Series Data Mining In Maintenance Of Ship

In the era of knowledge economy, the competition is increasingly intense. How to accessknowledge effectively has become the key to improve the competitiveness of enterprises. Theinformation technology of ship industry is also developing rapidly, in order to make the ship has thebest running state, the data of cycle maintenance must treat Seriously. However, the managementsystem only used to display the data has been unable to meet the user”s needs, due to theaccumulation of the massive maintenance data, the efficiency and ability of data analysis greatlylimited. Therefore, how to mining potential knowledge from the maintenance data quickly andaccurately is the urgent problem to be solved. So this paper combined data characteristics andrequirement of the system, proposed one kind OLAM system based on the time series data mining,focusing on mining the mass maintenance data rapidly, the main contents of this paper are as follows:(1) Given an OLAM system frame based on the time series data mining, which has presentationlayer, logic layer and data layer, and logic layer is the core parts of the system.(2) Proposed a kind of improvement OLAM system with working warehouse. Designed the datastructure and the data replacement algorithm of working warehouse, with the help of the warehouse,the I/O of server reduce, meet the needs of user interaction.(3) Designed a similarity search algorithm base on binary compression. Consider thecharacteristics of time series data and the problem that the huge data mining speed not suitable forWeb system. Designed a kind of algorithm which allow the user to set the compression ratio, andthrough the binary representation of time series data to achieve that accelerate speed of data miningwithout loss of the of time series, combined with the skyline index, achieves processingmultidimensional time series data efficiently.(4) Completed the design of OLAM system base on time series data mining and development ofmain function. Focus on the implementation details of the working warehouse and the binarycompression similarity search algorithm.The results of this research has been the application in a ship management system. The practiceindicated that the ability of analysis mass data improved greatly, and able to meet the needs of userinteraction. The accuracy of analysis results is greatly improved, and assist ship maintenance work.