python - Linear transformation for noisy data in python

Question

I have a dataset that you see below. The data is pretty noisy, but there is a clear linear trend that goes up and to the right. I'd like to transform the data with y = m * x to make the lines horizontal. Essentially, I'd like to do a regression on the orange lines to pull out the slope, but I don't know how to extract the different linear clusters. Is there a good method for transforming data like this? I'm using python/pandas/numpy.

Answer 1

看起来您会想要尝试对橙色点进行聚类。一些聚类方法将处理并行集群。我可能会从DBSCAN开始。

有关集群的更多信息，请查看此 scikit-learn 页面上的教程。你的情况有点像这里的第四行：

如果您提供您的数据，我希望有几个人会查看它。