Cross-Browser Incompatibilities Classification Layout: A comparative study between different models

Abstract

When the same web application is rendered in different browsers, inconsistencies detected in the layout or behavior of pages are known as (XBIs Cross Browser Incompatibilities). Currently, there are different classification models in the literature for the identification and automatic correction of XBIs. These models have evolved with the aim of reducing false positives and negatives. This paper proposes to compare these different models, focusing on those that use the classification of layout XBIs, through machine learning algorithms. There is still no paper in the literature to compare them, identifying their main advantages and disadvantages. This paper consists of an experiment that compares the results of models and presents metrics that allow to affirm how effective they are, aiming also to bring important information as contributions to propose future works regarding the evolution of the explored models. The result of the experiment is the metric of F-Score. For this metric, the higher values imply greater efficiency in detecting incompatibilities between browsers, and the C5.0 10 iterations - X configuration obtained the best result in the experiment.