lakerwsl / sstmetric-manuscript-code

This is the repository archiving code for the manuscript -- Self-Supervised Metric Learning in Multi-View Data: A Downstream Task Perspective

SSTMetric-Manuscript-Code

This is the repository archiving code for the manuscript -- Self-Supervised Metric Learning in Multi-View Data: A Downstream Task Perspective

File Introduction

The code provides all functions and scripts necessary to reproduce the results of the numerical experiments in the manuscript. In particular, the R scripts are divided into two collections: Simulation and RealData. In Simulation, the code includes four R scripts to reproduce the results in Section 6.1:

• sampleidentification.R for Table 3
• twosample.R for Figure 3
• kmeans.R for Table 4
• knn.R for Figure 4

In RealData, the code includes two R scripts to reproduce the results in Section 6.2:

• MNIST.R for the left part of Table 5
• Fashion_MNIST.R for the right part of Table 5

In these R scripts, data are generated or accessed by R functions. All results can be reproduced if the random seeds are chosen as in the R scripts.