A location prediction method and device based on a deep belief network

Abstract

The embodiment of the present invention provides a location prediction method and device based on a deep belief network, which relates to the field of location prediction technology, and solves the problem of only using the coordinates of the location in the prior art, using the Markov model for location prediction, and only using the location For the dimension of coordinates, the dimensions considered are not comprehensive enough, which affects the accuracy of the predicted position. Wherein the position prediction method based on the depth belief network comprises: using the multiple features as the input of the depth belief network, utilizing the depth belief network to perform position prediction, and using the output value of the depth belief network as the neuron corresponding to 1 The predicted position is predetermined to be the position to be reached by the user after the position of the point of interest; wherein, the deep belief network is to use samples including the same type of features as the multiple features as a training set, and the depth The belief network is trained. This improves the accuracy of the predicted location.

Description

technical field [0001] The present invention relates to the technical field of location prediction, in particular to a location prediction method and device based on a deep belief network. Background technique [0002] At present, the user's historical movement trajectory is used to predict the location after the historical movement trajectory, so as to provide services for users based on geographic location. For example, users often appear in places such as gyms and sports grounds. Combining the frequency and time of users going to gyms and sports grounds, the user's exercise habits can be inferred, and then the possible location of the user can be predicted, such as the playground. This can provide the user with navigation from the current location to the playground, or make recommendations about the location of sports equipment in the playground. [0003] For the problem of position prediction, the common method is the Markov model Markov Model, HMM (Hidden Markov Model,...

AI Technical Summary

Problems solved by technology

That is to say, only using the coordinates of the position and using the Markov model for position prediction, only using the dimension of the coordinates of the position, the dimensions considered are not comprehensive enough, which will affect the accuracy of the predicted position

Images