• A statistical (probabilistic) classifier: Predicts the probability of a given tuple belonging to a particular class 预测给定元组属于特定类别的概率
• Foundation: Based on Bayes’ theorem. Bayes was a mid-18th century monk (apparently).
  
• Performance: Comparable accuracy performance to decision tree and neural network classifiers
• Computational performance is much enhanced by assuming class-conditional independence, in which case the method is called Naive Bayes. 通过假设类条件独立性，计算性能大大提高，在这种情况下，该方法被称为朴素贝叶斯。
• Incremental: Each training example can incrementally contribute to the classification probabilities, so this allows adapting over time to gradual or incremental changes in (labelled) training data. 增量:每个训练例子都可以对分类概率做出增量贡献，因此这允许随着时间的推移适应(标记的)训练数据的逐渐或增量变化。
• It is not really possible to humanly interpret the results (i.e. it is known as a “black box” method), although it’s relationship to its training data is straightforward to understand. 虽然它与训练数据的关系很容易理解，但它实际上不可能被人理解(即它被称为“黑箱”方法)。