Statistics

Describe the issue linked to the documentation

Documentation should be changed to reflect that one-vs-rest is possible.

x = np.array([0,1,2,3,4,5,3,3,5,5,5,7,7,2])
x = x.reshape(-1,1)#this has 1 feature, therefore reshaping properly

y = [0,0,0,1,0,2,1,1,2,2,2,3,3,0] #note: y has multiple classes.


model = SVC(gamma = "auto", decision_function_shape="ovr")
model.fit(x,y)
p

Collection of follow-ups to #5827. These can/should be broken out into individual PRs. Many are relatively straightforward and would make a good first PR.

General

• Documentation (none was added in original PR).
• Release notes.
• Example notebook.
• Double-check how sm.tsa.arima.ARIMA works with fix_params (it should fail except when the fit method is statespace

See:
gonum/gonum#845 (comment)

The currently implemented version of the horseshoe distribution is not the parameterization that most ML papers use. This limits the ease of use of this as, for example, a prior in a tfp.layers.KLDivergenceAddLoss or in tfp.layers.DenseReparameterization. The regularized horseshoe would also be useful as an implemented distribution.

The alternative parameterization is shown here:
https://www.

Since the default output is meant to be human-readable, would it make sense to add thousands separators to make the output more easily readable?

For example, the data is (3.8,4.5,4.6,4.7,4.9)
while I'm using tech.tablesaw.aggregate.AggregateFunctions.percentile function, the 90th percentile is 4.9, however, if the percentile function supports linear interpolation, the 90th percentile should be 4.82, which is adopted by most other programming languages.

When we show data for a metric, we currently don't include the current day's worth of data. For users just getting set up, they may only have events from today, and want to test out if the query is working, and by excluding events from 'today', they can't see results.

TODO:

• In packages/back-end/src/services/experiments.ts on line 329, instead of using the current date as the value