lsorber / neo-ls-svm Goto Github PK

Ideas:

1. Allow the user to supply a target y and then apply a separating affine transform before running One-Class LS-SVM so that similarity is measured in y-separating feature space instead of the given feature space.
2. Perhaps use the resulting OC-LS-SVM (1) to fit a robust LS-SVM by reweighting or excluding samples.

References:

1. Least squares one-class support vector machine
2. One-class LS-SVM with zero leave-one-out error

Hi,

thanks for an awesome package. I wonder what it would take to add this model into polars api. Similar to polars-ols.

Ideas:

1. Fit a NeoLSSVM on y(t+1), then fit a new NeoLSSVM on y(t+2) | ŷ(t+1), and so on. This is more expensive than training a single t+1 model, but has the advantage that there's a direct estimate for the confidence interval of ŷ(t+k) with conformal prediction of the k'th model. Another advantage is that this method does not need 'future values of past covariates' as darts does.
2. Automatically determine the forecast horizon k by stopping when the confidence intervals are no longer meaningful.
3. Initially, only support past covariates and no future covariates. Later, add automatic future covariates based on the dataset's time series index using Time Axes Encoders. And after that maybe allow users to supply their own future covariates.
4. To keep things simple initially, don't allow supplying a new target to forecast a fitted model on (i.e. predict does not accept X or y). Instead, users can fit a new model if there's new data.