Abstract: Investment estimate prediction model in traditional transmission line（TTL）project has problems

such as large deviation from the actual cost and low efficiency of estimation management. Thus，a novel investment

estimate prediction model in TTL project was presented with assistance of principal component analysis and back

propagation（PCA-BP）neural network. Firstly，parameters that affect the investment in TTL projects were applied

as the initial input variables，which were dimensionally reduced with the help of PCA to simplify the complexity of

the input data. Then，used relevance pruning algorithm to optimize number of BP neural network nodes to further

improve the speed and accuracy of the proposed algorithm. Finally，an example study was carried out using data

from the estimated investment budget for transmission line projects of Hebei Electric Power Company from

January 2018 to January 2020 as a sample. The results show that the prediction accuracy of the designed PCA-BP

neural network-based probabilistic prediction model is 70% and 29% higher than that of the traditional support

vector machines（SVM）and BP neural network methods respectively，and have faster convergence and significant

engineering application value.

      Key words: principal component analysis（PCA）；back propagation（BP）neural network；transmission line；

investment budget