[MEO]Application Research of Naive Bayes Classification Algorithm in Weather Website

EasyChair Preprint 5685, version 1

Abstract

The update of the weather website travel products provides tourists with a reference to the weather conditions of the destination. However, due to various reasons, the weather website travel products cannot be updated on time. It is necessary to manually monitor whether it is updated. If it is not updated, it must be manually updated manually, which undoubtedly increases the business. The burden of personnel, in the intelligent era, It need to find a solution that saves time and effort to solve this problem. The Naive Bayes algorithm is widely used because of its advantages such as high classification accuracy and simple model. For this purpose, the Naive Bayes classification prediction algorithm combined with Python crawler is used to update the forecast of tourism products on the weather website. The algorithm combines the weather website's historical update data mining in the past month to calculate the a priori probability, and then calculates the classification result based on the Python program to capture the data of the day. By recording 16 sample data sets in the future, this model is used for calculation and analysis. 15 pieces of data conform to the results of the model calculation classification, and the accuracy rate reaches 93.7%. The results show that the high accuracy of algorithm classification prediction can remind business personnel in time, better guarantee the timely update of tourism products, thereby improving the efficiency of business personnel, and providing practical application reference value for the automation of meteorological service business.

Keyphrases: Crawler, Naive Bayes, classification prediction, meteorological, weather website

@booklet{EasyChair:5685,
  author    = {Chaoning Li and Liang Chen and Shenghong Wu and Yunyin Mo and Liying Chen},
  title     = {[MEO]Application Research of Naive Bayes Classification Algorithm in Weather Website},
  howpublished = {EasyChair Preprint 5685},
  year      = {EasyChair, 2021}}