FeatureWrap

Import FeatureWrap

To import FeatureWrap model use:

>>> from TINTOlib.featureWrap import FeatureWrap
>>> model = FeatureWrap()

When creating the FeatureWrap class, some parameters can be modified. The parameters are:

Parameters	Description	Default value	Valid values
`problem`	The type of problem, this will define how the images are grouped.	‘supervised’	[‘supervised’, ‘unsupervised’, ‘regression’]
`size`	The width and height of the final image, in pixels (rows x columns).	[8,8]	[int, int]
`bins`	The number of bins or intervals used for grouping numeric data	10	int
`zoom`	Multiplication factor that determines the size of the saved image relative to the original size. Values greater than 1 will increase the size of the saved image proportionally.	1	int
`verbose`	Show in terminal the execution.	False	[True, False]

Code example:

>>> model = FeatureWrap(size=[10,10], bins=20)

All the parameters that aren’t expecifically setted will have their default values.

FeatureWrap has the following functions:

Function

Description

Output

saveHyperparameters(filename)

Allows to save the defined parameters (scale, fea_dist_method, image_dist_method….).

.pkl file with the configuration

loadHyperparameters(filename)

Load FeatureWrap configuration previously saved with saveHyperparameters(filename)

filename: .pkl file path

generateImages(data, folder)

Generate one image per instance and group by class in different folder

data: path of the CSV or pandas dataframe

folder: path of the folder to save results

Folders with synthetic images

Paper: https://doi.org/10.1007/978-3-319-70139-4_87

Code Repository: