Spark - MLlib 线性回归截距和权重 NaN [英] Spark - MLlib linear regression intercept and weight NaN

问题描述

我尝试使用一些自定义数据在 Spark 上构建回归模型，截距和权重始终为 nan.这是我的数据:

I have trying to build a regression model on Spark using some custom data and the intercept and weights are always nan. This is my data:

data = [LabeledPoint(0.0, [27022.0]), LabeledPoint(1.0, [27077.0]), LabeledPoint(2.0, [27327.0]), LabeledPoint(3.0, [27127.0])]

输出:

(weights=[nan], intercept=nan)  

但是，如果我使用这个数据集(取自 Spark 示例)，它会返回一个非 nan 权重和截距.

However, if I use this dataset (taken from Spark examples), it returns a non nan weight and intercept.

data = [LabeledPoint(0.0, [0.0]), LabeledPoint(1.0, [1.0]), LabeledPoint(3.0, [2.0]),LabeledPoint(2.0, [3.0])]

输出:

(weights=[0.798729902914], intercept=0.3027117101297481) 

这是我当前的代码

model = LinearRegressionWithSGD.train(sc.parallelize(data), intercept=True)

我错过了什么吗?是不是因为我的数据上的数字那么大?这是我第一次使用 MLlib，所以我可能会遗漏一些细节.

Am I missing something? Is it because the numbers on my data are that big? It is my first time using MLlib so I might be missing some details.

谢谢

推荐答案

MLlib 线性回归基于 SGD，因此您需要调整迭代和步长，请参阅 https://spark.apache.org/docs/latest/mllib-optimization.html.

MLlib linear regression is SGD based therefore you need to tweak iterations and step size, see https://spark.apache.org/docs/latest/mllib-optimization.html.

我像这样尝试了你的自定义数据，我得到了一些结果(在 Scala 中):

I tried your custom data like this and I got some results (in scala):

val numIterations = 20
val model = LinearRegressionWithSGD.train(sc.parallelize(data), numIterations)

这篇关于Spark - MLlib 线性回归截距和权重 NaN的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！