说明：  本报告对 朴素贝叶斯模型及线性判别分析、二次判别分析 进行系统测试 “生成模型”是机器学习中监督学习方法的一类。与“判别模型”学习决 策函数和条件概率不同，生成模型主要学习的是联合概率分布??(??,??)。本 文中，我们从朴素贝叶斯算法入手，分析比较了几种常见的生成模型（包 括线性判别分析和二次判别分析）应用于多因子选股的异同，希望对本领 域的投资者产生有实用意义的参考价值。
(This report gives a systematic test of naive Bayesian model, linear discriminant analysis and two discriminant analysis "Generation model" is a kind of supervised learning method in machine learning. Learning from "discriminant model" The strategy function and the conditional probability are different. The generation model mainly studies the joint probability distribution. book In this paper, we analyze and compare several common generation models (packages) from the naive Bayes algorithm. The linear discriminant analysis and the two discriminant analysis are applied to the similarities and differences of multi factor stock selection. The investors of the domain have a useful reference value.)