Research On Taxi Behavior Detection Method Based On Large Data Mining

Distributed processing platform has high efficiency when handling big data. With Application of distributed processing platform, large taxi GPS data are analyzed, and the results can be used to efficiently analyze the data characteristics of taxi GPS. For taxi GPS data processing, the clustering analysis algorithm is a relatively efficient algorithm. The ability of the improved K-means algorithm has been much improved when processing GPS data.The improved K-means algorithm is a hybrid algorithm; the choose of the initial clustering center is more accurate, and correspondingly, the results of the data progress are also accurate when using hybrid algorithm. The improve K-means algorithm can be progressed in two parts: firstly, DBSCAN clustering algorithm is used to cluster the big data, acquiring the initial clustering and the initial cluster center, then analyzing center datasets K; secondly, the results obtained based on the K value of clustering results need to be processed with K-means algorithm to get the final optimized processing result.After GPS data preprocessing, the data processing model is established under the hybrid algorithm. Taxis’ peak period, load rate duration and other issues are analyzed with the model. For long taxi passenger analysis, the main comparative of the passenger business days and rest days when long-duration distribution, and the distribution of the peak of the peak level of passenger, in the study of load rate, respectively, from the time the load rate and mileage rate study load. Finally, according to the results to achieve practical application, distributed according to the taxi pick-hot time, spatial distribution and spatial distribution of the passengers on the establishment of walk the shortest distance taxi stand models.This research has very important practical value. Through the analysis of the characteristics of taxi GPS data, we can understand the travel law of urban residents, and help to ease urban traffic congestion. In addition, through the analysis of the no-load rate and other issues, we can understand whether the city taxi quantity is appropriate or not in order to achieve the reasonable use of urban resources and to avoid the waste of resources.