摘要

A dimension reduction model is proposed in this paper. The proposed model can be used to transform the unit's biding curve of power producers of a market into a one-dimensional feature vector, which reflects the changes of the bidding curve quite conformably. As a result, the unit's biding curves can be classified by using cluster analysis. Through clustering calculations of the bidding Curves of Zhejiang electricity market, it is shown that the bidding curve can be classified accurately and efficiently by using a 20-segment model and the squared sum of deviations.

全文