The Application Of Data Mining Technology In Microsoft Customer Support Service Systems

Microsoft Corporation is leading company in personal computer software industry.It is now the largest computer software supplier in the world. To Microsoft Customer Support Service department, on one hand, the leadership team is facing the requirement of raising the return of investment from service perspective while increase overall competitiveness of Microsoft products and services. On the other hand, the major business operation rules are made based on subjective past experience despite the massive archive of raw support data technical support incidents with the growth of software technical support business.Modern main stream Data Mining models are quite mature and systematic. However, it is still difficult to apply existing Data Mining models onto software technical support, which incorporates both technology and service, due to the number metrics and metrics types inconsistent. Massive data analysises are still based on linear analysis for continuous single metrics. The problem of combining continuous and descrete data for a unified relativity analysis and the problem of how to analyze distribution status of huge amount of data utilizing modern database technology advantage are still remain unsolved.This Data Mining research was carried out in the above context. The article utilized the modern Data Mining technologies through careful study and application. Appropriate Data pre-processing and suitable Data Mining models like Decision Tree and Box Plot were used to turn raw data regarding customer support quality and working efficiency into practical business rules through data extracting using Data Warehouse technologies against historical data in Microsoft software technical support systems. Such business rule information were then transformed into employee and manager friendly form through visual intelligent client. This project has provided convincing scientific proof for process refine in Software Techncial Support business.The most important innovative work in this article includes re-architecturing Data Mining oriented data cube using modern Data Warehouse technology together with understanding of business model. Descretlize continuous metrics data using business relative Data Generalization mechanisms so that relative analysis model, Decision Tree model, can apply. The previous work ensures that ID3 algorithm can be applied onto multiple dementional data of software technical support. The other innovative part in this article is to implement efficient algorithm for analyzing service labor metrics distribution within different customer satisfaction sample space using the Quartile distrituion analysis and Boxplot model. This algorithm incorporates the Database advantage over data sorting and fast look-up.During the implementation phase, the system also creatively used database connection, auto-calculate and data chart features in Microsoft Excel software to accomondate business logics, data querying and visualized presentation into one single easy to maintain/upgrade intelligent client.With the landing of this system, it confirmed the fact that it is possible to discover business improvement knowledge and information from real world business data to benefit management and frontline engineers through proper user interface together with appropriate Data Mining models. This project has setup a solid foundation for informationlize and scientifictize business analysis for Microsoft Customer Support Service business.