Recent Releases of pca

pca - 2.10.0

  • All verbosity messages are now handled by the logger.

If you experience any difficulties, let me know in the issues!

- Jupyter Notebook
Published by erdogant 6 months ago

pca - 2.0.8

  • Update for warning get_cmapwhen using matplotlib >= 3.10 (Issue #59)
  • Fix for error in biplot that sometimes occurred (issue #58)

- Jupyter Notebook
Published by erdogant about 1 year ago

pca - 2.0.7

  • Pass the verbose parameter into scatterd issue #56
  • Fix type issue #52

- Jupyter Notebook
Published by erdogant over 1 year ago

pca - v2.0.5

  • Fix when fontsize is set to 0. A new release of matplotlib causes some warnings/errors.

- Jupyter Notebook
Published by erdogant over 2 years ago

pca - v2.0.4

  • Can handle standardized verbose messages as input parameter.

- Jupyter Notebook
Published by erdogant over 2 years ago

pca - v2.0.3

  • plt.show removed for plotting
  • bug fix in case 3D plot is requested but number of PCs is <3
  • Examples are imported using datazets library
  • Code cleaning and pep8 styling

- Jupyter Notebook
Published by erdogant over 2 years ago

pca - v2.0.2

  • added parameter grid and opaque_type to scatter.

- Jupyter Notebook
Published by erdogant over 2 years ago

pca - v2.0.1

  • Updated arrow coloring in the biplot.
  • More information can be found here

- Jupyter Notebook
Published by erdogant over 2 years ago

pca - v2.0.0

Developing new functionalities is really cool. However, when making incremental improvements over time, the code complexity also gradually increases. I took the time to refactor the entire plotting part. When using this version, you likely need to rename some input parameters in your code. But it is worth it because the plots became even more beautiful!

  • scattering is now performed in scatterd library
  • Many input parameters for plotting are aligned to the scatter functionality of matplotlib.
  • for plotting, some parameters such as textlabel are removed because were redundant.
  • for plotting, the parameter y is renamed into labels
  • it is now possible to add density and gradient into the plots and keeping the plot look nice
  • Changing the ordering of the density layer is possible (on top or below)
  • Fix for 3d plot and the positioning of the text labels
  • High improvements in plotting speed when having many data points!
  • updated documentation, docstrings and readme

- Jupyter Notebook
Published by erdogant over 2 years ago

pca - v1.9.2

  • Removed wget as a dependency
  • Added functionality to specify markers per sample
  • Improved functionality to customize alpha transparency per sample
  • Improved functionality to customize size per sample
  • Updated docstrings

- Jupyter Notebook
Published by erdogant almost 3 years ago

pca - v1.9.1

  • fix for multiple testing
  • Added parameter ax to plots

- Jupyter Notebook
Published by erdogant almost 3 years ago

pca - v1.9.0

  • set default std=3 wich is more common for outlier detection
  • Multiple test corrections for the hotelling t2 test
  • multipletestsis set in the predict function and not during initialization anymore.
  • y_probais the corrected Pvalue. Prawis the uncorrected Pvalue in the output dataframe

- Jupyter Notebook
Published by erdogant almost 3 years ago

pca - v1.8.6

  • Font color inherits the arrow color (default).
  • Font colors can be adjusted in the plots.
  • Sizes of the scatter can be adjusted with parameter s.
  • Colors of the scatter ben be adjusted with parameter c.
  • fig can be given as an input parameter to make iterative changes to the plot.

Examples can be found here.

- Jupyter Notebook
Published by erdogant almost 3 years ago

pca - v1.8.5

  • sklearn changed into scikit-learn in both requirements and setup file.

- Jupyter Notebook
Published by erdogant about 3 years ago

pca - v1.8.4

  • Fix for overlapping annotations in biplot issue #33
  • Updated docstrings.
  • Set the same order for input parameters for the plotting functions.

- Jupyter Notebook
Published by erdogant about 3 years ago

pca - v1.8.3

Fix for big arrowheads.

- Jupyter Notebook
Published by erdogant over 3 years ago

pca - v1.8.2

  • Added boolean to choose whether to update outlier parameters or not.
  • Example can be found here.

- Jupyter Notebook
Published by erdogant over 3 years ago

pca - v1.8.1

  • Detect and plot outliers for new unseen transformed samples

- Jupyter Notebook
Published by erdogant over 3 years ago

pca - v1.8.0

  • added parameter method to change between pca, sparsepca and truncsvd.
  • Fix for trunc_svd.

Examples: ``` model = pca(method='sparsepca') model = pca(method='truncsvd') model = pca(method='pca')

```

- Jupyter Notebook
Published by erdogant over 3 years ago

pca - v1.7.2

  • Fix for not showing scatter in case of biplot using cmap=None

- Jupyter Notebook
Published by erdogant over 3 years ago

pca - v1.7.1

  • Improved speed in plotting in case of having thousands of samples by setting parameter label=None

model.biplot(label=None)

- Jupyter Notebook
Published by erdogant over 3 years ago

pca - v1.7.0

  • Density coloring implemented with the gradientparameter.

In this example, the cmap=Set1will be used to color the class labels. The coloring will have a continuous scale towards the borders.

pca.scatter(cmap='Set1', gradient='#ffffff')

- Jupyter Notebook
Published by erdogant almost 4 years ago

pca - v1.6.4

  • Title can be changed in the figures
  • Fix when normalizing out PCs by the conversion of dataframe into a numpy array.

- Jupyter Notebook
Published by erdogant almost 4 years ago

pca - 1.6.3

  • Fix for newer versions of Python related to dict_items. Thanks @tgy!
  • Fix in case no explained variance is detected.

- Jupyter Notebook
Published by erdogant almost 4 years ago

pca - 1.6.2

  • The .plot functionality will show the total explained variance across the components when selecting on explained variance: n_components<1
  • docstring updates
  • Some code cleaning

- Jupyter Notebook
Published by erdogant almost 4 years ago

pca - 1.6.1

  • Fixes for biplot when choosing different PC to plot.
  • New parameter for biplot, to color the arrow: model.biplot(color_arrow='g')

- Jupyter Notebook
Published by erdogant almost 4 years ago

pca - 1.6.0

  • Added parameter in biplotand biplot3dto choose the PC axes.

``` fig, ax = model.biplot(PC=[0, 1]) fig, ax = model.biplot3d(PC=[0, 1, 2])

```

- Jupyter Notebook
Published by erdogant almost 4 years ago

pca - 1.5.5

  • added doi

- Jupyter Notebook
Published by erdogant about 4 years ago

pca - 1.5.4

  • Fix in case PCs are selected based on percentage explained variance. In cases when new samples needed to be transformed with the existing model, an error occurred.
  • Licence added to the setup file

- Jupyter Notebook
Published by erdogant about 4 years ago

pca - 1.5.3

  • Fix for plot in case only 1PC is detected.

- Jupyter Notebook
Published by erdogant about 4 years ago

pca - 1.5.2

  • Added functionality to train/infer mode for outlier detection in Hotelling's T2 and SPE #15 and #16
  • Added functionality to use lists as input for the model #17

- Jupyter Notebook
Published by erdogant over 4 years ago

pca - 1.5.1

  • Detection of outliers is now optional. If you do not want to detect outliers use the following during initialiation: detectoutliers=None. The detection of outliers can now be set by: detectoutliers='ht2', 'spe'

- Jupyter Notebook
Published by erdogant over 4 years ago

pca - 1.5.0

Many improvements thanks to Daniel Scott!

  • Added alpha_transparency option & tests
  • Fixed syntax in Readme
  • Fix default value in docstring
  • Added alpha_transparency argument & tests for all relevant plots
  • Added pytest github workflow
  • Fix pytest yaml syntax
  • Added miniconda pandas to github workflow so that test suite can run
  • Added requirements to pytest
  • Removed explicit reference to tests
  • Removed macosx support from tests. Colormap not supported on osx"
  • Removed explicit reference to tests
  • Moved tests dir
  • Initialized test directory
  • Added wget to dependencies
  • Improved commenting in test coverage
  • pytest fix for slow CI on windows
  • pytest fix for slow CI on windows (contd)
  • Added numpy installation to conda
  • pytest fix for slow CI builds
  • Disable testing for windows (only linux)

- Jupyter Notebook
Published by erdogant over 4 years ago

pca - 1.4.0

  • Transforming unseen data is now possible with the "transform" function

- Jupyter Notebook
Published by erdogant almost 5 years ago

pca - 1.3.0

  • visible status can be set in plots

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.2.0

  • Change colormap in biplot using the cmap parameter
  • Some pep8 styling.
  • Bug fix that resulted in an error in very unusual cases.

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.1.2

  • Bug fix: ellipse was removed before returning

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.1.1

  • Direct plotting from spe_dmodx is possible with the showfig parameter

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.1.0

  • outlier detection with spe/DmodX
  • plotting SPE/DmodX
  • results of spe/dmodx in key 'outliers'
  • Hotelling T2 test and SPE are separate boolean parameters.
  • Boolean parameter outlier is removed.
  • Update in examples

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.9

  • Outlier detection incorporated using hotelling T2 test and fishers method.
  • Outliers in plots
  • Outliers docstrings
  • Outliers example in readme file

Have fun with outliers!

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.8

  • loadings are shown in biplot
  • the output of ['topfeat'] contains additional columns where the selected loadings are shown.
    • bug fix in 3d plot
  • additional parameter "label" to plot labels in scatter/biplot

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.7

  • fix for length arrows in plot
  • fix for number of features to plot in biplot

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.6

  • fix for huge plots

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.5

  • fix in biplot
  • fix in biplot3d (now with correct arrows)
  • fix in determining top features

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.4

PC naming was a bit to much changed. ITs back now.

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.3

  • input variable sparse_data changed into onehot
  • example add

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.2

  • name "coef" used for loadings.

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.1

  • fix for biplot plotting weights

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 1.0.0

  • Huge changes because pca is refactored to work with classes!

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 0.1.7

  • x-axis ticks can be partially numbered

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 0.1.6

  • Some fixes in plotting: x-axis numbering starts with 1 instead of 0.
  • Docstings updated with output parameters
  • cleaning files

- Jupyter Notebook
Published by erdogant over 5 years ago

pca - 0.1.5

  • parallel processing for sparse_datasets
  • input parameter sparse changed into sparse_data

- Jupyter Notebook
Published by erdogant almost 6 years ago

pca - 0.1.4

  • normalize data (default True)
  • bug fixes
  • unit tests

- Jupyter Notebook
Published by erdogant almost 6 years ago

pca -

  • Sparse matrix support with TruncatedSVD
  • Sparse data support with SparsePCA
  • width and heigth changed in figsize()
  • Code cleaning

- Jupyter Notebook
Published by erdogant almost 6 years ago

pca - minor updates

  • fit outputs 'X' instead of 'pc'
  • plot_explainedvar changed into plot

- Jupyter Notebook
Published by erdogant almost 6 years ago

pca - First release

- Jupyter Notebook
Published by erdogant almost 6 years ago