This is a simple package to fit a circle to points in a plane.

Currently, there are 3 methods implemented (Kasa, Pratt, Taubin).

13 Likes

To expand the package a bit I want to to include some methods provided by `StatsBase`

like `dof`

, `residuals`

, `coef`

, `rss`

.

I was thinking of introducing a new return struct which can handle those methods.

Is there a common naming scheme or procedure on how to do handle for results?

To also implement `fit`

and `fit!`

I think I need a struct to represent the model itself, not only the results. Are there any good examples on how to implement a `RegressionModel`

?

The new version is currently developed in the statsmodel branch.

1 Like

Just registered version 0.2, which now uses a StatsBase.RegressionModel like interface.