modAL
modAL copied to clipboard
modAL-python
SVR regerssion
Hello, when I use SVR for regression task, it will prompt: 'SVR' object has no attribute 'predict_proba' I think it should be that there is no query_strategy. So how can I design a query that matches regression?
Hi! I don't know how to estimate uncertainties with SVR, so I can't help you with that. However, if you can use Gaussian processes instead of SVR, the given query strategy would work.
The error message is correct since the query strategy tries to call the predict_proba() method of the regressor. This is not available for SVR.
I'm glad to hear from you. I've used a number of other tasks and query strategies in the categorization task, besides the SVR.Next I want to create some regression tasks and select the commission's query strategy.But I didn't find any relevant samples. Could you give me some advice?I would appreciate it. Thank you！
The only example I have is the active regression with Gaussian processes: https://modal-python.readthedocs.io/en/latest/content/examples/active_regression.html
I am unfamiliar with any active regression tasks with SVR, so I cannot help you with this :(
OK. Thanks again! I will try other methods!
Hi @zt823793279 did you find any query strategy. i am also trying for regression task and i really don't know what strategy will work. Can you please help me in this regard please
I have same problem.....how to apply modAL to regression problem rather than Gaussian process?
