This repository contains an open-source implementation of Semi-Supervised Locally Linear Embedding (SSLLE) by Yang et al. (2006).
The code base for SSLLE is stored in 1_sslle_implementation.
In order to compute an SSLLE embedding, simply use the perform_sslle
function defined in fun_perform_sslle.R.
The easiest way to set up all required packages and source files is by simply
running 0_run_setup.R first.
All additional code concerning analyses and visualization is stored under
0_seminar/1_code.
0_utils contains source files (general utility functions and functions
for visualization purposes) required to run the scripts besides the core SSLLE
implementation.
1_scripts contains code required to produce the sensitivity analysis
(1_run_sensitivity_analysis.R) and corresponding visualization
(2_run_visualization_experiments.R).
3_run_figures.R produces all figures used for report and presentation.
Prior to running these files, 0_run_setup.R must be executed to set up
all packages and source the required function files.
Lastly, 2_data contains several temporary data files produced by the above
scripts as well as the raw input data needed to build the world data set.
... compute an SSLLE embedding, simply use perform_sslle from
1_sslle_implementation.
... reproduce all analyses and figures of the seminar report/presentation,
execute all scripts in 0_seminar/1_code/1_scripts in the given order.