zhangys11 / cla

toolkit for classifiability analysis

cla (classifiability analysis)

A unified classifiability analysis framework based on meta-learner and its application in spectroscopic profiling data [J]. Applied Intelligence, 2021, doi: 10.1007/s10489-021-02810-8

pyCLAMs: An integrated Python toolkit for classifiability analysis [J]. SoftwareX, Volume 18, June 2022, 101007, doi: 10.1016/j.softx.2022.101007

Warning

Since 0.3.x, we have reorganized the package structure. Any upper app should be revised accordingly.
Since 1.0.0, we stopped pyCLAMs and switch to cla.

Installation

pip install cla (pyCLAMs for versions under 1.0.0)
pip install rpy2
Install the R runtime and the ECol library (https://github.com/lpfgarcia/ECoL).

Run 'install.packages("ECoL")' in R. It will take very long time. You must wait for the installation to complete.
Sometimes, you may want to change the CRAN mirror. Under the "Packages" menu, click "Set CRAN Mirror".
After installation, you can check by R command 'installed.packages()'.

How to use

Download the sample dataset from the /data folder Use the following sample code to use the package:

  # import clams # (for versions < 1.0.0)  
  from cla import metrics # (for versions > 1.0.0)  

  # load the dataset or generate a toy dataset by X,y = mvg(md = 2)
  df = pd.read_csv('sample.csv')
  X = np.array(df.iloc[:,:-1]) # skip first and last cols
  y = np.array(df.iloc[:,-1])

  # get all metrics
  metrics.get_metrics(X,y) # Return a dictionary of all metrics

  # get metrics as JSON
  metrics.get_json(X,y)

  # get an html report and display in Jupyter notebook
  from IPython.display import display, HTML
  display(HTML(metrics.get_html(X,y)))

Start the web GUI

1. python -m cla.gui.run
2. Open http://localhost:5005/ in your browser.

3. A ready-to-use online demo is http://spacs.brahma.pub/research/CLA

---

Metrics and functions added since the original publication

1. metrics

classification.Mean_KLD - mean KLD (Kullback-Leibler divergence) between ground truth and predicted one-hot encodings
correlation.r2 - R2, the R-squared effect size
test.CHISQ, test.CHISQ.log10, test.CHISQ.CHI2 - Chi-squared test
classification.McNemar, classification.McNemar.CHI2 - McNemar test on the groud-truth and classifier's prediction
classification.SVM.Margin - the linear-SVC's margin width
test.student, test.student.min, test.student.min.log10, test.student.T, test.student.T.max
test.KW, test.KW.min, test.KW.min.log10, test.KW.H, test.KW.H.max
test.Median, test.Median.min, test.Median.min.log10, test.Median.CHI2, test.Median.CHI2.max

2. refactor

Integrate some existing packages and reorganize the package structure.

module	sub-module	description	standalone pypi package (if any)	publication
cla	cla.metrics	Provides various classifiability analysis metrics.	pyCLAMs	pyCLAMs: An integrated Python toolkit for classifiability analysis [J]. SoftwareX, Volume 18, June 2022, 101007, doi: 10.1016/j.softx.2022.101007
	cla.unify	Provide a method for unifying multiple atom metrics.	N/A	A unified classifiability analysis framework based on meta-learner and its application in spectroscopic profiling data [J]. Applied Intelligence, 2021, doi: 10.1007/s10489-021-02810-8
	cla.vis	Data visualization and plotting functions.	N/A	N/A
	cla.gui	Provide a user-friendly GUI.	wCLAMs	N/A

cla's Issues

missing dependencies in the setup

I wanted to let you know that when I installed cla from PyPI there were a few required packages that were not specified by the setup or requirements files:

1. pyNNRW
2. h5py
3. qsi_tk (which on my system required cmake)