初探PHP神经网络和机器学习--phpml

/**
 * Created by PhpStorm.
 * User: ctexthuang
 * Date: 2018/8/27
 * Time: 上午09:56
 */
require 'vendor/autoload.php';

//基于贝叶斯定理
use Phpml\Classification\NaiveBayes;

$samples = [
    [176,70],
    [180,80],
    [161,45],
    [163,47],
    [186,86],
    [165,49],
];

//通过一些特征，告诉我们机器这一组数据的结构   男生1  女生2
$labels = [2,1,2,2,1,2];

//采用贝叶斯来分类实现机器学习的分类
$classifier = new NaiveBayes;

$classifier->train($samples,$labels);

print_r($classifier->predict([[176,90]]));

/**
 * Created by PhpStorm.
 * User: ctexthuang
 * Date: 2018/8/27
 * Time: 上午10:13
 */
//大数据分析 ，先有数据源
$samples = [ //数据越多，结果越精准， 才有参考价值
    ['接口开发平台','swoole','分布式集群'],
    ['laravel','swoole','设计模式'],
    ['接口开发平台','swoole','rpc接口开发'],
    ['接口开发平台','swoole','框架内核'],
    ['游戏服务器开发','swoole','php7'],
    ['laravel','swoole','分布式集群'],
    ['mysql性能优化','接口平台开发','laravel'],
    ['redis分布式','网站负载均衡','mongodb'],
    ['handopp集群','机器学习','linux'],
];

$labels = [];

//预测算法，挖掘数据关联得到我们的结果可以用来消费市场分析，猜测顾客消费习惯

//算法
$associator = new Apriori($support = 0.3); //预测 -> 数据关联挖掘
$associator->train($samples,$labels);
//支持度 置信度
$res =$associator->predict(['swoole']);

var_dump($res);exit;