SMOGN’s documentation!

smogn module

class smogn.SMOGN(threshold=0.9, over_sampling_ratio=0.1, under_sampling_ratio=1.0, k=10, relevanse_base=0.5, pert=0.02, metric='minkowski')

Bases: object

Class performe over and under sampling for regression data.

This Object is implementation of SMOGN

Attribuites:

threshold (float):

threshold of rare example. [0, 1]

over_sampling_ratio (float):

ratio of over sample rare example. [0, 1]

under_sampling_ratio (float):

ratio of under sample normal example. [0, 1]

k (int):

number of nearest neighbors

relevanse_base (float):

base parameter of relevance_fn

pert (float):

pertubation parameter of gaussian noise

metric (str):

metric of distance.

relevances (np.array):

relevance values

fit_transform(X, target_column)

Parameters

X (pd.DataFrame) – training examples

Returns

new training examples

Return type

newX (pd.DataFrame)

relevance_fn(y, k=0.5, eps=1e-08)

calcuate relevance

Parameters

• y – np.array target examples

• k – float relevance base

• esp – float small value to avoid zero divition

Returns: relevance values: np.array

relevance values of y

Indices and tables

• Index

• Module Index

• Search Page

Created by Convergence Lab. https://www.convergence-lab.com