Based On Improved K Nearest Neighbor Classification Of It Asset Management System Development

HP solution for IT asset management (ITAM) could very well help enterprises to manage all IT asserts, to achieve the purpose of controlling costs and reducing risk. As the core part of solution, HP Discovery and Dependency Mapping Inventory (DDMI) can automatically discover and identify all installed software and hardware facilities in the target machine, and is an important data source of ITAM. But the application software identification method of DDMI is very complex, the calculation of process is huge, the data structure is messy, and still has a high probability recognition error. Furthermore, the labor costs will be increased with enterprises'grown in future because of much manual operation. So, it is important to find a way to reduce the complexity of the process, increase the degree of automation, and much accurate and comprehensive to satisfy the user's recognition requirements.First, the status of ITAM in HP is discussed and the necessity of application with KNN is also indicated. Based on it, a discussion about main functions and processes of automatic discovering and automatic identifying of the ITAM system is performed. Then the system architecture with MVC pattern is analyzed, and the core subsystems such as IT asset discovering and IT asset identifying are designed in detail. By analyzing the disadvantage of the engine, KNN classification is introduced as a basic implementation methods and the improved KNN algorithm with essential vector is employed in recognition engine. It obtains the K candidates by the first KNN with essential vector and then uses the second KNN with them. Experiments show that the improved algorithm EV-KNN can greatly cut the training sample and reduce the computational complexity. The high efficiency and accuracy are also shown in classification. The realization of EV-KNN is mainly analyzed. Finally the defects of the engine and the future research directions are summarized.