README.md

In amirms/GeLaToLab: GELATOLAB

GeLaToLab

A collection of visualization tools and algorithms for data analysis and machine learning focused on source code analysis.

Installation

# Install devtools from CRAN
install.packages("devtools")

# Or the development version from GitHub:
# install.packages("devtools")
devtools::install_github("amirms/GeLaToLab")

Multi-view Learning

Multi-view Clustering

Function perform.clustering in file clusterValidate.R performs evaluation of single and multi-view clustering for each project. It finds the best kernel functions + parameters for each view, and uses the same choices to perform multi-view clustering for different methods.

For clustering, complete hierarchical cluster analysis is used, and path difference (PD) is used to measure the performance.

Multi-view CF-based Recommendation System

Function perform.prediction in file recommenderValidate.R perform single-view and multi-view evaluation of CF-based recommendations for each project using a nested k-fold cross validation. The default setup is a 10-fold nested cross-validation.

It first finds the best kernel parameters for each view, followed by evaluating three multi-view learning approaches to perform CF-based recommendation.

The scores used to measure the performance are: ROC AUC, PR AUC, and max F1 scores.

Uni- and Cross-modal search

In crossModalRetrieval.R file, function findCrossModal performs the experiment for uni-modal and cross-modal retrieval based on the best kernel parameters for each view.

amirms/GeLaToLab documentation built on May 12, 2019, 2:36 a.m.