I am investigating a dataset and have found the outcome has three, or maybe four, distinct groups of data with positive correlation. enter image description here

I would like to fit two lines to classify my data points by which of the three trends they are part of. With the equations for these classifiers, I would then classify each point into high, middle, or low so I can examine correlations in the covariates.

My first thought would be to use the perceptron, but I would like to do this without going through and labeling each point by hand. What other approaches can I take to do this? Perhaps something regression based?

I realize I could fit trends by hand, but would like an automatic and reusable method so I can easily change model parameters.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.