| With the development of China’s socioeconomic and the growth of electricity demand, the investment in power grid construction increasing rapidly, power companies take a huge investment pressure and heavy construction tasks. However, China’s current investment levels are still rising, and the investment return is also at a low level compared to other developed countries, which greatly limits the development of power grid project. Therefore, a comprehensive investment management to improve the level of power transmission project is in need, investment forecasting as a core project of the early stages of decision-making is the key to control investment, Therefore, it has important practical significance to study the power transmission project investment forecasting methods.This paper analyses the limitations of existing forecasting methods and models based on a lot of relevant literature. Preliminary selects the investment forecasts factors through analyzing the impact from two levels of transmission project about the external environment and internal features, and as a practical basis for investment prediction model; Analyzing the primary factors using rough set-kernel principal component, first uses rough set reduction redundancy factor, then uses nuclear principal component analysis to extract major influence component, applying it in support vector machine prediction model which were completed by PSO method, completing the construction of the transmission line project prediction model and giving a specific forecasting process; Combining a sample history of a region of transmission line engineering, chooses the related influencing factors and processes its data, using the constructed model for training and testing, confirms the validity of the model, and compares the improved model with the traditional model, the results show that the multilevel model predictions factor screening system has good adaptability and accuracy, the average relative error is reduced by 40%, and the relative error is less than 15%; Finally, proposes the development optimization recommendation of intelligent predictive through the problem of model building practices and lessons learned in the case of power transmission projects, provides a useful reference for other project investment forecasting. |
