LibSampling

A Library of Sampling Methods for 2D Scatterplots

Introduction

LibSampling is a python-based library of sampling methods for 2D scatterplots, which includes:

• random sampling
• density biased sampling [1]
• blue noise sampling [2]
• farthest point sampling [3]
• non-uniform sampling [4]
• Z-order sampling [5]
• SVD based sampling [6]
• hashmap based sampling [7]
• outlier biased random sampling [8]
• outlier biased blue noise sampling [9]
• outlier biased density based sampling [9]
• multi-class blue noise sampling [10]
• multi-view Z-order sampling [5]
• recursive subdivision based sampling [11]

Our goal is to facilitate the use of the popular sampling methods in visualization, graphics, and data mining. LibSampling provides a simple python interface where users can easily apply an appropriate sampling method to their data. The characteristics of the sampling methods and the evaluation results of our study are listed as follows:

Requirement

• cffi==1.11.2
• scipy==1.0.0
• numpy==1.13.3+mkl
• Flask==0.10.1
• Werkzeug==0.14.1
• matplotlib==2.1.1
• scikit_learn==0.19.1

Download LibSampling

The current release (Version 1.0, April 2020) of LibSampling can be obtained by directly cloning this repository.

Quick Start

Below is the example code to call a sampling method, where n and m are the sizes of the input and output data (sampling result), respectively.

from sampling.Sampler import *
from sampling.SamplingMethods import *
sampler = Sampler() # Initialize the Sampler object
sampler.set_data(data, labels)
# data: an (n*2) numpy array indicating the point coordinates
# labels: an (n*1) numpy array indicating the class labels with intergers 0, 1, 2, ...
# or None if you use single-class sampling methods
sampler.set_sampling_method(RandomSampling, sampling_rate=0.5)
indices = sampler.get_samples_idx()
# indices: an (m*1) numpy array indicating the indices of the sampling result
sampled_data, sampled_labels = sampler.get_samples()
# sampled_data: an (m*2) numpy array indicating the sampled point coordinates
# sampled_labels: an (m*2) numpy array indicating the class labels of the sampled points

A more complete example of how to use LibSampling is provided in example_sampling.py :

import numpy as np
from sampling.Sampler import *
from sampling.SamplingMethods import *

print("run Sampler.py")

# Simple synthetic data
points = np.random.random((10000, 2)) # Generated data, the input data should be a numpy array with the shape (n, 2)
categories = np.random.randint(0, 10, 10000) # Generated label, multi-class sampling method would consider the label information as an reason to select or not select an item. It would be a np.zeros(n) as default.

# Datasets used in our study
all_data = np.load(os.path.join('data', 'abalone.npz'))
points, categories = all_data['positions'], all_data['labels']

print(points.shape, categories.shape)

sampler = Sampler()

sampler.set_data(points, categories) # For single-class sampling methods like random sampling, categories is not needed to be provided
sampling_method = RandomSampling # You can choose your desired sampling method.
rs_args = {
    'sampling_rate': 0.3 # You can set the sampling ratio and other specific params for different sampling methods here.
}

sampler.set_sampling_method(sampling_method, **rs_args) # Set Random Sampling for the sampler with necessary params
sampled_point, sampled_category = sampler.get_samples() # Get the sampling result

print("Random sampling result:")
print(sampled_point, sampled_category)

sampling_method = OutlierBiasedRandomSampling
outlier_score = np.sum(np.abs(points - 0.5), axis=1)
obrs_args = {
    'sampling_rate': 0.5, # You can set the specific params for different sampling methods here, e.g., sampling rate
    'outlier_score': outlier_score # The default outlier_score will be determined by the class purity if you do not pass your own outlier_score to outlier biased sampling methods
}

sampler.set_sampling_method(sampling_method, **obrs_args) # Set Outlier Biased Random Sampling for the sampler with necessary params
sampled_point, sampled_category = sampler.get_samples() # Get the sampling result

print("Outlier biased random sampling result:")
print(sampled_point, sampled_category)

sampling_method = RecursiveSubdivisionBasedSampling
rsbs_args = { # This sampling method do not need sampling rate as input
	'canvas_width': 1600,
	'canvas_height': 900,
	'grid_width': 20,
	'threshold': 0.02,
	'occupied_space_ratio': 0.02,
	'backtracking_depth': 4
}
sampler.set_sampling_method(sampling_method, **rsbs_args)
sampled_point, sampled_category = sampler.get_samples() # Get the sampling result

print("Recursive subdivision based sampling result:")
print(sampled_point, sampled_category)

Datasets

The folder data contains the datasets in the npz format used in our study.

Direct link: https://github.com/thu-vis/libsampling/tree/master/data

You can obtain them by unzipping the npz file:

# Datasets used in our study
all_data = np.load(os.path.join('data', 'abalone.npz'))
points, categories = all_data['positions'], all_data['labels']

As most of them are high-dimensional datasets, we transform them into 2D data using t-SNE and normalize them to [0, 1] × [0, 1]. The basic information of these datasets is as follows:

The projection results of the full datasets are shown in the following figure, where colors encode classes.

Trial data

The folder trial data contains the trial data of our evaluation.

You can obtain our analysis results by directly run process.py.

Additional Information

Please cite our work as:

@article{Yuan2021Evaluation, 
    author={J. Yuan and S. Xiang and J. Xia and L. Yu and S. Liu}, 
    journal={IEEE Transactions on Visualization and Computer Graphics}, 
    title={Evaluation of Sampling Methods for Scatterplots}, 
    year={2021}, 
    volume={27}, 
    number={1}, 
    pages={1--1}
}

J. Yuan, S. Xiang, J. Xia, L. Yu, and S. Liu. Evaluation of Sampling Methods for Scatterplots. IEEE Transactions on Visualization and Computer Graphics, 27(2):1720–1730, 2021.

Acknowledgments

The lib sampling/RSBS.dll for Recursive subdivision based sampling is derived from the source code at the homepage of [11]. (Link: https://github.com/Ideas-Laboratory/RecursiveSubdivision-basedSampling)

References

[1] C. R. Palmer and C. Faloutsos. Density biased sampling: An improved method for data mining and clustering. In Proceedings of the ACM SIG- MOD International Conference on Management of Data, pages 82–92, 2000.

[2] R. L. Cook. Stochastic sampling in computer graphics. ACM Trans. Graph., 5(1):51–72, 1986.

[3] M. Berger, K. McDonough, and L. M. Seversky. cite2vec: Citation-driven document exploration via word embeddings. IEEE transactions on visualization and computer graphics, 23(1):691–700, 2016.

[4] E. Bertini and G.Santucci. By chance is not enough: preserving relative density through non-uniform sampling. In Proceedings of the Eighth International Conference on Information Visualisation, pages 622–629. IEEE, 2004.

[5] R. Hu, T. Sha, O. VanKaick, O. Deussen, and H. Huang. Data sampling in multi-view and multi-class scatterplots via set cover optimization. IEEE Transactions on Visualization and Computer Graphics, 26(1):739–748, 2020.

[6] P. Joia, F. Petronetto, and L. Nonato. Uncovering representative groups in multidimensional projections. Computer Graphics Forum, 34(3):281–290, 2015.

[7] S. Cheng, W. Xu, and K. Mueller. ColorMapND: A data-driven approach and tool for mapping multivariate data to color. IEEE Transactions on Visualization and Computer Graphics, 25(2):1361–1377, 2019.

[8] S. Liu, J. Xiao, J. Liu, X. Wang, J. Wu, and J. Zhu. Visual diagnosis of tree boosting methods. IEEE transactions on visualization and computer graphics, 24(1):163–173, 2017.

[9] S. Xiang, X. Ye, J. Xia, J. Wu, Y. Chen, and S. Liu. Interactivecorrection of mislabeled training data. In Proceedings of the IEEE Conference on Visual Analytics Science and Technology, pages 57–68, 2019.

[10] L. -Y. Wei. Multi-class blue noise sampling. ACM Transactions on Graphics, 29(4):79, 2010.

[11] X. Chen, T. Ge, J. Zhang, B. Chen, C. Fu, O. Deussen, and Y. Wang. A recursive subdivision technique for sampling multi-class scatterplots. IEEE Transactions on Visualization and Computer Graphics, 26(1):729– 738, 2020.

[12] D. Surendran. Swiss roll dataset. https://people.cs.uchicago.edu/~dinoj/manifold/swissroll.html, 2004. Last accessed 2019-04-30.

[13] A. Coraddu, L. Oneto, A. Ghio, S. Savio, D. Anguita, and M. Figari. Machine learning approaches for improving condition-based maintenance of naval propulsion plants. Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment, 230(1):136–153, 2014.

[14] B. A. Johnson and K. Iizuka. Integrating OpenStreetMap crowdsourced data and landsat time-series imagery for rapid land use/land cover (LULC) mapping: Case study of the laguna de bay area of the philippines. Applied Geography, 67:140–149, 2016.

[15] Y. Lecun, L. Bottou, Y. Bengio, and P. Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278– 2324, 1998.

[16] R. G. Andrzejak, K.Lehnertz, F. Mormann, C. Rieke, P. David, and C.E. Elger. Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state. Physical Review E, 64(6):061907, 2001.

[17] D. Dua and C. Graff. UCI machine learning repository, 2017.