Inhalation Particle Matter Forecasting Analysis Based On Chaos Theory In Chengdu

It is getting higher and higher regarding on environment quality, with the enhancement of people living standard. Additionally, atmospheric density of Inhalation Particle Matter (PM10) becomes more and more important.In view of this, this article uses the monitor data of the atmospheric density of Inhalation Particle Matter (PM10) of Chengdu in 2007, and obtains the time series of noise of the atmospheric density of Inhalation Particle Matter (PM10) of Chengdu, through the wavelet-denoise which is constituted by Mallat algorithm decomposition and restructuring. Additionally predicting the time series of noise has been carried on for the first time.At first, the frequency-power figure and the median period of the time series of noise of the atmospheric density of Inhalation Particle Matter (PM10) has been obtained, through the fast Fournier transformation first (Fast Fourier Transformation, FFT).Then, delay-time and time-window of the time series of noise has been known through C-C method .Additionally ,the proper number of inserting-dimension and the delay-time has been estimated.Based on the former, the biggest Lyapunov index of the time series of noise is 0.0357 is worked out through the small data quantity method, and the biggest Lyapunov index forecast model has carried on to the time series of noise.The result indicated:1)The biggest Lyapunov index forecast model based on the chaos theory can forecast density of Inhalation Particle Matter (PM10) of Chengdu accurately, and error is 21.75%, which is little more than network model that has dealt by B-P network optimization algorithm and RBF neural network model, much more accurate than the other models, the one of the high accurate models;2)It concludes the system is chaotic because of that the biggest Lyapunov index is positive number;3)The predict results have already demonstrated high erroneous and the precision change high dramatic when the predict time is nearness of the max forecast time .So the forecast time should be reduced as short as possible.Finally, the two aspects to increase the precision has been pointed out this article. First one is the massive practices and the improvement inserting dimension and the delay time, and the other is new model based on chaos theory besides the biggest Lyapunov index forecast model should be carried on to increase the accurate.