Method for predicting vehicle using behavior mode change of driver in real time

The invention relates to a method for predicting vehicle using behavior mode changes of a driver in real time. The method comprises the steps that historical data of personal travel of a driver are acquired in real time and preprocessed, and the historical data are vehicle use time and travel space position of the driver in the previous travel; the method comprises the following steps: normalizing historical data, dividing a travel map into grids, giving an index to each grid in the grids as spatial data, dividing time sequence data by day, counting the use time of an individual vehicle every day, taking the use time as time data, and obtaining the counted time sequence data; and inputting the time sequence data into a pre-constructed generalized likelihood ratio model, analyzing the vehicle use mode of the current period, and obtaining vehicle use mode change information of the driver in the current period. The private car use behaviors of a plurality of individuals are deduced in a probability distribution