Spatiotemporal Traffic Modeling based on Frequent Pattern Mining in Wireless Cellular Network

Prediction has been one of the major purposes in big data analytics, and has shown great impacts in many fields. In this paper, a frequent pattern mining method is proposed, called Frequent Pattern Mining-Matching (FPM Matching), for traffic prediction with spatiotemporal model. To demonstrate the effectiveness of spatiotemporal traffic modeling which based on FPM-Matching in wireless cellular network, extensive experiments of the wireless cellular network are conducted at two big cities in China. The statistics of the collected data are examined, including the correlation coefficients among the base stations, and the traffic distribution of all the base stations. Then, the 15-days data was divided into training sets and test sets. The temporal and spatial correlation of the base stations are found from the test sets and used as structured data to conduct pattern matching in prediction process. The FPM-Matching algorithm is proposed to use the frequent tree-pattern sets model to store frequent sequences, so that the frequent sequences build up the tree structure according to the hierarchical and node progressive way. In this case, each frequent sequence will have a position coordinate in the structure. In the process of prediction, largest matrix to achieve pattern matching is proposed. Through the frequent pattern sets of tree structure, it can determine the position of prediction mode quickly, obtaining the candidate sets. The evaluation results show that the FPM-Matching algorithm has the advantages of accuracy and effectiveness in prediction based on spatiotemporal modeling. FPM-Matching use the training set effectively to obtain stable model parameters and achieve accurate prediction of the traffic state. This is important for the base station's sleep strategy, bandwidth, power and resource block control. No matter it is from the point of time complexity or accuracy, the algorithm has obvious advantages and save resource reasonably.