_ML_

Machine learning is very important, and this is what I sum up from this project: the model needs to be iteratively modified, 
that is, back propagation. We should improve the model through data from different dimensions and stages.
Depending on the dimension of the feature vectors you extract, after repeated verification of the different number of dimensions 
in this project, it is concluded that the best effect can be achieved when the dimension is 3, and the gambling company is 
an individual.
Now it's time to collect data from N gambling companies. A can provide false information, but BCDE can't provide false information 
at the same time in a game.
So the "three-board model" I'm exploring is based on three boards to determine a stable geometry, that is, 
three gambling companies can determine the outcome of a game.

三木板模型算法论文地址：http://wanfangdata.com.cn 搜索“基于机器学习的预测算法模型及其在环评领域的应用”
注：在一场比赛开始前，菠菜公司90%是知道赛果的，他们已经有精算师分析基本面、伤病等信息，我只要站在他们的肩膀上进一步分析就可以了。（无需再分析基本面等信息）
当时FBP整个项目学习都是靠以下进行学习，在此感谢作者！希望大家参考后也能建立属于自己的机器学习项目！
PS：内容风趣幽默，关键是零基础！
http://www.captainbed.net/blog-acredjb