As a simple and efficient classification algorithm, Naive Bayes (NB) algorithm has been applied in the field of harassing calls recognition, But in reality, it is difficult to satisfy its attribute independence hypothesis. The Weighted Bayes (WB) is an improved algorithm that assigns different weights for different feature attributes to reduce the impact of attribute independence hypothesis on classification results. Different calculation methods of weights will have a great influence on the

more »

... influence on the classification performance. By fruit fly optimization algorithm (FOA) can optimize the weight of attributes, which obtain an Improved Weighted Bayes algorithm (IWB) to improve the performance and accuracy of the recognition of harassing calls. Simulation results show that compared with NB algorithm and WB, IWB algorithm can apparently improve the effect of the recognition of harassing calls, and the recognition accuracy is increased by 3.8% as well.