Machine learning usage methods and scenarios:
Bayesian classifier - Churn analysis: The lost users can be used as a sample, which includes the single use time, cost, active interaction, etc., for churn early warning;
Bayesian classifier + text mining - spam filtering, web page text classification, automatic analysis of comments (good, medium, and bad): first prepare samples, segment words, make TF-IDF matrix, and then use Bayesian classification algorithm to calculate Which category should it belong to
Wavelet Analysis - OCR, Image Recognition, Speech Recognition
Decision tree - computer game (with position standardization, position evaluation scoring function, chess score learning, etc.), because it is temporarily unavailable at work, this part is not the focus