Abstract: Real estate price is formed in the circumstance of many factors since real estate market is a complex system. Empirical mode decomposition, an effective tool to deal with nonlinear and non-stationary data, is applied to analyze the time series of real estate price and reveal inherent characteristics. Weekly price series of the new residential housing traded in Hangzhou during the last four years is decomposed, and six intrinsic mode functions and one residue are obtained. The six IMFs are reorganized according to their characteristics. It is concluded that the time series of real estate price are constituted by three parts, i.e., long-term trend determined by economic fundamentals, low frequency vibration brought by major events, and random fluctuations caused by short-term market disequilibrium. Besides, it is revealed that new residential housing market in Hangzhou has a large cycle of three-year and also small cycles of 14 months and 7 months.

Key words: empirical mode decomposition (EMD), housing price analysis, periodic fluctuations, complex system, housing price forecast