Statistics.Sample.KernelDensity.Simple is deprecated?

[jsermeno]

Hi,

I wanted to learn more about this module's deprecation. I want to specify the lower and upper bounds for an estimate, where the lower and upper bounds extend beyond the range of the sample data. I was unable to get the results I want with Statistics.Sample.KernelDensity.kde_, but found that Statistics.Sample.KernelDensity.Simple.estimatePDF works perfectly. My question is, should kde_ be able to do everything that estimatePDF can? Is there something I need to do to the data I pass into kde_? Here's an example:

With the following test data:

let demoData = V.fromList $ concat $ zipWith replicate [5, 4, 3, 7] [0..3]

The following estimate:

-- xValues has 'n' values ranging from 'a' to 'b'.
Stats.estimatePDF Stats.gaussianKernel bw demoData (Points xs)

Returns an estimate like this:

While the following code:

Stats.kde_ n a b demoData

Returns an estimate like this:

Thanks

[Shimuuar]

@bos could you comment on this?

[luke-clifton]

This looks related to #46