The Analysis Of Manufacturing Enterprise Sales Based On Data Warehouse

Computer assistance decision system can help manufacturing enterprises integrating management of production, operation, design, manufacture. It can provide enterprises accurate and valid data information, and bring huge benefit. Therefore, it makes sense to research decision support technique applied to enterprise management. Decision support system is able to establish a production sales analysis system for manufacturing enterprise, acquire accurate, completed and timely status information, discover and analyze problems deeply in time, and offer a human-computer interface platform.This paper is facing product sales of a chemistry building materials group. It discusses On-line Analytical Process and Data Mining according to data warehouse technique.The main research content and result are as follows:(1) It studies how to expand the original system, transform only on line analytical process system to a system combinated on line analysis with data mining.(2) It discusses data warehouse design process arming at product sale analysis, designs the logic model and physics model of data warehouse, and contains something about extracting transformation and loading data for data warehouse.(3) It shows the components of entity, fact, and measure for analyzing sales on line with Micorstrategy, creats different product sale reports, receives intuitionist result, and is convenient for enterprise production management.(4) It makes use of Matlab to statistically regress the key product sales trend with time, putts forword a regress model, then works out it and analyzes it.(5) It brings forward simple classification arithmetic, and founds the K-Means cluster model of centres classification. It develops program to carry out the two kinds of classification arithmetic with C++ Builder6.0.The research result of the thesis has applied to the group, which shows this system is able to analyze product sales from different point, such as produce sales region gap, sales trend forecast, and checking filiales achievement. It is practiced actually.The result is testifed well.