Research Of Data Mining Technology Used In Management Of Engineering Cost For Power Construction

In recent years, with the rapid development of the national economy, the current capacity of power grid was put to a severe test. The power reform aims at building a strong national grid in the future to support rapid economic development. According to statistics, in 2007 the national grid construction investment amounted to 202.5 billion Yuan, and in "11th Five-Year Plan" period, the cost of construction of power grids will be increased to 1 trillion Yuan, and such an enormous amount of investment will inevitably bring about great increase of grid construction projects, at the same time, it will brought a large number of issues about the investment management.The power construction projects are implemented the responsibility system of construction Project Corporation. The project corporation is responsible for project planning, financing, construction in the whole process and assumes corresponding risks. This project management made the corporate level to a very high demand. Cost management is the most important part of project management for its planning and financial control functions. In the past, controlling of project cost often focus on the consumption of raw materials, and whether construction spending coincides with the design value during the construction period, this method is rationality a certain extent, but the control of overall cost does not significantly effective.This paper use data mining technologies to discover technical and economic knowledge in cost management. All the analysis processes are accord with the standard flow of data mining, preprocessing data firstly and choosing core algorithms to build models later. This paper builds two models to assistant project corporation to calculate project cost quickly: engineering. cost forecast model based on fuzzy neural network and engineering cost evaluation model based on sensitivity analysis. The first model, can be used to calculate the engineering cost rapidly with little parameters in the initial phase of construction, and can help the corporate to earn enough time for investment planning, financing and project operation. The second model can choose the most important attributes to certain unit project cost. The corporate can use these two models to implement fast evaluation and other management functions.