1, simple and clear
There are many ways to regularize the model in classification, and there is no need to worry about the correlation between features like naive Bayesian classifier. They can easily handle the interaction between features and are nonparametric, so there is no need to worry about whether outliers or data are linearly separable.
2, easy to explain and explain
When classifying, the selected attributes or features are systematized in a certain order to form a scientific and reasonable classification system. This helps to explain and explain, and the classification should meet the needs of the continuous development and change of things.
3, high accuracy
Classification is a logical method to classify things with certain similarity or similar characteristics into an uncertain set by comparing the similarities between things. Can organize and systematize a large number of complex materials; Discover and master the universal law of the development of things.
Introduction to Bayesian classification
Bayesian classification technology plays an important role in many classification technologies, and it also belongs to the category of statistical classification. This is an irregular classification method. Bayesian classification technology trains the classified sample subset, learns and induces the classification function, and uses the trained classifier to classify the unclassified data.
Bayesian method is widely used in classification problems because it gives the optimal solution to minimize the error in theory. On the basis of Bayesian method, Bayesian network method is proposed. Naive Bayesian classification assumes that one attribute has an independent influence on a given classification. This assumption is called conditional independence, which greatly simplifies the calculation required for classification.