Naive Bayesian algorithm classification model with local attribute weighted based on KNN

Abstract

In order to better optimize and configure public transport resources, we will find the rules of the bus lines different people taking, and predict which bus lines different people choosing. To solve this problem, this paper proposed a new Naive Bayesian algorithm classification model with local attribute weighting based on K-nearest neighbor algorithm. In the process of algorithm calculated, we used the K-nearest neighbor algorithm to find the K neighbors to be classified, then calculated the probability of each attribute in K neighbors as the weight of the attribute. Later, we put the weight of attribute into Naive Bayesian classification process, which makes the classification model more realistic and predicted the label. We used K-nearest neighbor, decision tree and Gaussian naive Bayesian algorithm as control group. Experiments were carried out on bus historical data in one city. The results show that the model has high accuracy in the bus line selection prediction.