GeneralizedLinearAlgorithm in Spark MLLib
SparkMllib涉及到的算法
- Classification
- Linear Support Vector Machines (SVMs)
- Logistic regression
- Regression
- Linear least squares, Lasso, and ridge regression
- Streaming linear regression
GeneralizedLinearAlgorithm

GLA,通用线性算法,作为通用回归算法(regression)和分类算法(classification)的抽象算法,run函数中实现了算法的流程,并最终产生通用线性模型。抽象算法流程主要包括:addIntercept, useFeatureScaling, 调用optimizer计算参数等。其中optimizer是抽象类,具体的线性算法需要指定具体的optimizer实现类。【模板模式——子类的训练大多调用了GLA中的run函数】

进一步地,optimizer是算法的核心,也是迭代算法的逻辑所在。具体类包括梯度下降等算法。这些优化算法又可以抽象并归纳出两个模块:梯度计算器(Gradient)和参数更新(Updater)。
综上,对于某一个具体的线性算法时主要一个优化器,而优化器是通过选择梯度计算器和参数更新器的组合来得到。【bridge pattern】
当前版本中GLA有多个子类,覆盖了多个分类和回归算法:
- LassoWithSGD (org.apache.spark.mllib.regression)
- LinearRegressionWithSGD (org.apache.spark.mllib.regression)
- RidgeRegressionWithSGD (org.apache.spark.mllib.regression)
- LogisticRegressionWithLBFGS (org.apache.spark.mllib.classification)
- LogisticRegressionWithSGD (org.apache.spark.mllib.classification)
- SVMWithSGD (org.apache.spark.mllib.classification)
例如,LinearRegressionWithSGD算法,对应的Optimizer是GradientDecent,对应的梯度计算器是LeaseSquareGradent,x位置的梯度$$\vec{\triangledown} = (\vec{x}^T\cdot\vec{w} - y)\cdot\vec{x}$$,对应的更新器为SimpleUpdater,更新算法是$$\vec{w} =: \vec{w} - \alpha / \sqrt{iter}\cdot\vec{\triangledown}/m$$
GeneralizedLinearModel
Algorithm will create a model by calling:
def createModel(...): M
**GeneralizedLinearModel **(GLM) represents a model trained using GeneralizedLinearAlgorithm. GLMs consist of a weight vector and an intercept.
Parameters:
- weights - Weights computed for every feature.
- intercept - Intercept computed for this model.
基本上每个线性算法都对应到一个线性模型: - LassoModel
- LinearRegressionModel
- LogisticRegressionModel
- RidgeRegressionModel
- SVMModel
GradientDecent梯度下降法
梯度下降法对convex函数必然能求解。
function [theta, J_history] = gradientDescent(X, y, theta, alpha, num_iters)
%GRADIENTDESCENT Performs gradient descent to learn theta
% theta = GRADIENTDESCENT(X, y, theta, alpha, num_iters) updates theta by
% taking num_iters gradient steps with learning rate alpha
% Initialize some useful values
m = length(y); % number of training examples
J_history = zeros(num_iters, 1);
for iter = 1:num_iters
% ====================== YOUR CODE HERE ======================
% Instructions: Perform a single gradient step on the parameter vector
% theta.
%
% Hint: While debugging, it can be useful to print out the values
% of the cost function (computeCost) and gradient here.
%
theta = theta - alpha * X' * (X * theta - y) / m;
% ============================================================
% Save the cost J in every iteration
J_history(iter) = computeCost(X, y, theta);
end


代码中可以发现多处使用了梯度下降法,包括ann,LR/SVM分类算法,Lasso/Linear/Ridge回归等算法。
主要构造函数:
class GradientDescent private[spark] (
private var gradient: Gradient,
private var updater: Updater)
extends Optimizer with Logging {
private var stepSize: Double = 1.0
private var numIterations: Int = 100
private var regParam: Double = 0.0
private var miniBatchFraction: Double = 1.0
private var convergenceTol: Double = 0.001
其中gradient为计算梯度的公式,updater为根据梯度的值去更新权重的公式。除了这两个参数之外,还有算法训练过程中的一些技术参数。
算法的基本参数设定之后便,再给定训练样本,便可以训练得到模型权重:
def optimize(
data: RDD[(Double, Vector)],
initialWeights: Vector): Vector = {
val (weights, _) = GradientDescent.runMiniBatchSGD(
data,
gradient,
updater,
stepSize,
numIterations,
regParam,
miniBatchFraction,
initialWeights,
convergenceTol)
weights
}
=> 调用迭代梯度下降法
def runMiniBatchSGD(data: RDD[(Double, Vector)],
gradient: Gradient,
updater: Updater,
stepSize: Double,
numIterations: Int,
regParam: Double,
miniBatchFraction: Double,
initialWeights: Vector,
convergenceTol: Double): (Vector, Array[Double])
Run stochastic gradient descent (SGD) in parallel using mini batches. In each iteration, we sample a subset (fraction miniBatchFraction) of the total data in order to compute a gradient estimate. Sampling, and averaging the subgradients over this subset is performed using one standard spark map-reduce in each iteration.
Parameters:
- data - Input data for SGD. RDD of the set of data examples, each of the form (label, [feature values]).
- gradient - Gradient object (used to compute the gradient of the loss function of one single data example)
- updater - Updater function to actually perform a gradient step in a given direction.
- stepSize - initial step size for the first step
- numIterations - number of iterations that SGD should be run.
- regParam - regularization parameter
- miniBatchFraction - fraction of the input data set that should be used for one iteration of SGD. Default value 1.0.
- initialWeights - initial weights for model training.
- convergenceTol - Minibatch iteration will end before numIterations if the relative difference between the current weight and the previous weight is less than this value. In measuring convergence, L2 norm is calculated. Default value 0.001. Must be between 0.0 and 1.0 inclusively.
Returns:
A tuple containing two elements. The first element is a column matrix containing weights for every feature, and the second element is an array containing the stochastic loss computed for every iteration.
迭代梯度下降法 (SGD)
上述分析可以发现梯度下降法采用的实现法法是SGD,该算法对模型进行迭代更新,非常适合分布式计算,算法流程:
- Sample a subset of data or just use all the data;
- [MAPPER] for each entry, calculate gradient and model loss by
Gradient$$(g^{(i)}, l^{(i)}) := \cal{Gradient}(x^{(i)}, y^{(i)}, \theta)$$; - [AGGREGATION] sum of gradient = $$\sum_ig^{(i)}$$, sum of loss = $$\sum_i l^{(i)}$$, counter = $$\sum_i 1$$;
Updater$$\theta' := \cal{Update}(\theta, \sum_ig^{(i)} / \sum_i 1, \alpha, ...)$$;- if $$||\theta'-\theta||_2 / ||\theta'||_2 < convergenceTol$$ finish, else $$\theta := \theta'$$, repeat from step 1.
各线性算法的G/U组合
| Algorithm | optimizer | Gradient | updater |
|---|---|---|---|
| LinearRegressionWithSGD | GradientDescent | LeastSquares | Simple |
| LassoWithSGD | GradientDescent | LeastSquares | L1Updater |
| RidgeRegressionWithSGD | GradientDescent | LeastSquares | SquaredL2 |
| LogisticRegressionWithSGD | GradientDescent | Logistic | SquaredL2 |
| LogisticRegressionWithLBFGS | LBFGS | Logistic | SquaredL2 |
| SVMWithSGD | GradientDescent | Hinge | SquaredL2 |
GradientDecent以及SGD算法的流程在上面已经描述,接下来具体分析各种梯度计算和权重更新算法:
LeaseSquareGradent
lost function:
\]
\]
SimpleUpdater
A simple updater for gradient descent without any regularization.
更新算法是$$\vec{w} := \vec{w} - \alpha / \sqrt{iter}\cdot\vec{\triangledown}$$,
其中$$ \vec{\triangledown} = \sum_i{\vec{\triangledown_i} / \sum_i {i}}$$
算法迭代过程中整体的步长会随着迭代次数而逐渐减小,

SquaredL2Updater
L2 regularized problems:
L = 1/n ||X \cdot w - y||^2 + R(w),\\
\text{then:} \frac{dL}{dw}=\vec{\triangledown} + \lambda w\\
w:=w' - \alpha \cdot \frac{dL}{dw}\\
\ \ \ = (1-\alpha \lambda)\vec{w} - \alpha / \sqrt{iter}\cdot\vec{\triangledown}\]
与LinearRegressionWithSGD比较,RidgeRegressionWithSGD的优化器依然是LeastSquaresGradient,训练样本计算Gradient额算法也相同,不通点是在给定Gradient结果基础上更新权重时,old权重需要额外乘以一个系数。
GeneralizedLinearAlgorithm in Spark MLLib的更多相关文章
- Spark MLlib之线性回归源代码分析
1.理论基础 线性回归(Linear Regression)问题属于监督学习(Supervised Learning)范畴,又称分类(Classification)或归纳学习(Inductive Le ...
- Spark MLlib - LFW
val path = "/usr/data/lfw-a/*" val rdd = sc.wholeTextFiles(path) val first = rdd.first pri ...
- 《Spark MLlib机器学习实践》内容简介、目录
http://product.dangdang.com/23829918.html Spark作为新兴的.应用范围最为广泛的大数据处理开源框架引起了广泛的关注,它吸引了大量程序设计和开发人员进行相 ...
- Spark MLlib 之 Basic Statistics
Spark MLlib提供了一些基本的统计学的算法,下面主要说明一下: 1.Summary statistics 对于RDD[Vector]类型,Spark MLlib提供了colStats的统计方法 ...
- Spark MLlib Data Type
MLlib 支持存放在单机上的本地向量和矩阵,也支持通过多个RDD实现的分布式矩阵.因此MLlib的数据类型主要分为两大类:一个是本地单机向量:另一个是分布式矩阵.下面分别介绍一下这两大类都有哪些类型 ...
- Spark MLlib - Decision Tree源码分析
http://spark.apache.org/docs/latest/mllib-decision-tree.html 以决策树作为开始,因为简单,而且也比较容易用到,当前的boosting或ran ...
- Spark入门实战系列--8.Spark MLlib(上)--机器学习及SparkMLlib简介
[注]该系列文章以及使用到安装包/测试数据 可以在<倾情大奉送--Spark入门实战系列>获取 .机器学习概念 1.1 机器学习的定义 在维基百科上对机器学习提出以下几种定义: l“机器学 ...
- Spark入门实战系列--8.Spark MLlib(下)--机器学习库SparkMLlib实战
[注]该系列文章以及使用到安装包/测试数据 可以在<倾情大奉送--Spark入门实战系列>获取 .MLlib实例 1.1 聚类实例 1.1.1 算法说明 聚类(Cluster analys ...
- spark mllib配置pom.xml错误 Multiple markers at this line Could not transfer artifact net.sf.opencsv:opencsv:jar:2.3 from/to central (https://repo.maven.apache.org/maven2): repo.maven.apache.org
刚刚spark mllib,在maven repository网站http://mvnrepository.com/中查询mllib后得到相关库的最新dependence为: <dependen ...
随机推荐
- lazarus的动态方法和虚拟方法
动态方法和虚拟方法在delphi里面分别表示: 动态方法 当需要调用父类.祖先类的被覆盖方法的时候,是查找继承树,当找到,就调用.减少了VMT占用,但调用慢一些. 虚拟方法 和动态方法不同的是,记录了 ...
- xcode资源管理
1. 在根目录放图片. UIImageView *image = [[UIImageView alloc] initWithImage:[UIImage imageNamed:@"ok.pn ...
- Sophus链接错误
错误指示如下: CMakeFiles/run_vo.dir/run_vo.cpp.o: In function `main': run_vo.cpp:(.text.startup+0x1086): u ...
- samtools flagstat
samtools flagstat命令简介: 统计输入文件的相关数据并将这些数据输出至屏幕显示.每一项统计数据都由两部分组成,分别是QC pass和QC failed,表示通过QC的reads数据量和 ...
- 使用jackson美化输出json/xml
转载:http://www.cnblogs.com/xiwang/ 如何使用jackson美化输出json/xml 1.美化POJO序列化xml 下面将POJO列化为xml并打印. Person pe ...
- iOS知识基础篇 static
static关键字的作用 一.隐藏 通过static修饰的函数或者变量,在该文件中,所有位于这条语句之后的函数都可以访问,而其他文件中的方法和函数则不行: 二.静态变量 类方法不可以访问实例变量(函 ...
- Linux+mysql+apache+php
1.1.1 所需软件 cmake ncourse mysql apr apr-util pcre apache php 1.1.2 解压缩软件 ...
- s4-6 二层交换
为什么需要二层交换? 有很多LAN,如何将它们连接起来? 可用网桥(bridges )将它们连接起来. 网桥工作在DLL层,通过检查MAC地址做出转发帧的决策 不会检查网络层,所以,IPv ...
- 如何制作一个自适应手机、电脑、ipad的网页方法总结
进入2015年,手机上网的用户已经越来越多,已经赶超PC端.随着2G.3G.4G.免费WIFI和无线基站的不断普及,越来越多的人开始使用手机上网. 移动设备正超过桌面设备,成为访问互联网的最常见终端. ...
- IP之ALTIOBUF仿真
这里实现了差分转单端的功能. 问题:差分信号的电平是怎样的?如果像平常一样不设置Pin Planner中的电平的话,编译会报错. 在Pin Planner中做了如下设置: `timescale 1 n ...