Today, learning punch, is a very simple model, naive Bayes method (naive bayes classifier)
Generally speaking is the Bayesian + Naive
By Bayesian probabilities to calculate the incident:
Then, Naive is the assumption that various factors are independent of each other.
In real life, factors that are often intrinsically linked. Such as: whether the rain today, considerations are: air pressure, humidity, temperature. In fact, these factors are intrinsically linked, but the model assumed that they are independent of each other, so called naive. In this way, the calculation is quite simple, and often predict the results are fairly good.
link:
https://pan.baidu.com/s/14HRsY341zzhfXSVWrdAGXg