DistrictDataLabs / yellowbrick

Visual analysis and diagnostic tools to facilitate machine learning model selection.
http://www.scikit-yb.org/
Apache License 2.0
4.3k stars 559 forks source link

Add description of parameter to docstring #1236

Closed lwgray closed 2 years ago

lwgray commented 2 years ago

@Haebuk had opened a PR to deal with this but closed it https://github.com/DistrictDataLabs/yellowbrick/pull/1233

We just need to add this line below to the KneeLocator docstring for y located in yellowbrick/utils/kneed.py

"A list of k scores corresponding to each value of k. The type of k scores are determined by the metric parameter from the KElbowVisualizer class. "

https://github.com/DistrictDataLabs/yellowbrick/blob/092c0ca25187b3cde9f608a1f7bc6d8c2b998f96/yellowbrick/utils/kneed.py#L61-L62

class KneeLocator(object):
    """
    Finds the "elbow" or "knee" which is a value corresponding to the point of maximum curvature
    in an elbow curve, using knee point detection algorithm. This point is accessible via the
    `knee` attribute.

    Parameters
    ----------
    x : list
       A list of k values representing the no. of clusters in KMeans Clustering algorithm.

    y : list
       A list of k scores corresponding to each value of k. The type of k scores are determined by the metric parameter from the KElbowVisualizer class.

    S : float, default: 1.0
       Sensitivity parameter that allows us to adjust how aggressive we want KneeLocator to
       be when detecting "knees" or "elbows".

    curve_nature : string, default: 'concave'
       A string that determines the nature of the elbow curve in which "knee" or "elbow" is
       to be found.

    curve_direction : string, default: 'increasing'
       A string that determines tha increasing or decreasing nature of the elbow curve in
       which "knee" or "elbow" is to be found.

    online : bool, default: False
        kneed will correct old knee points if True, will return first knee if False

    Notes
    -----
    The KneeLocator is implemented using the "knee point detection algorithm" which can be read at
    `<https://www1.icsi.berkeley.edu/~barath/papers/kneedle-simplex11.pdf>`
SangamSwadiK commented 2 years ago

@lwgray Can I please submit a PR on this ?

lwgray commented 2 years ago

Sure, make sure it is on the development branch.

And thanks for using yellowbrick