Evaluation-Framework

It is an evaluation framework for evaluating and comparing graph embedding techniques

Classification

Datasets used as gold standard

Dataset Semantic of classes Classes Size Source
Cities Living style 3 212 Mercer
AAUP Salary of professors 3 960 JSE
Forbes Agency income 3 1,585 Forbes
Albums Album popularity 2 1,600 Metacritic
Movies Movie popularity 2 2,000 Metacritic

Model and its configuration

Model Configuration
Naive Bayes -
C 4.5 decision tree -
k-NN k=3
Support Vector Machine C = {10^{-3}, 10^{-2}, 0.1, 1, 10, 10^2, 10^3}

Missing entities are simply ignored.

The results are calculated using stratified 10-fold cross-validation.

Output of the evaluation

Metric Range Optimum
Accuracy [0,1] Highest