Network Actual Traffic Prediction Algorithm Based on α-stable Distribution and Wavelet Transformation

In order to improve the prediction precision of wireless sensor network traffic, a new prediction algorithm (State Prediction algorithm based on α-stable distribution α, SP-α) is proposed, combined with α-stable distribution and wavelet transformation. The algorithm proposed first defines the characteristics of α-stable distribution and then gives the judge basis that obeys α-stable distribution. At the same time, it reduces the prediction error of the actual traffic by fusion of the prediction results of α-stable distribution with wavelet transformation. Finally, the paper thoroughly researches the key factors impacting on the new algorithm through simulations in OPNET and MATLAB. Compared with the performance of FARIMA model, the simulation results proved that SP-α algorithm has better adaptability.