A comparison of classification models for imbalanced datasets

Details

Supervisors

Saerens, Marco

Faculty

Louvain School of Management

Degree label

Master [120] en Ingénieur de gestion

Abstract

Many practical classification problems are imbalanced; i.e., at least one of the classes constitutes only a very small minority of the data. For such problems, the interest usually leans towards correct classification of the minor class. Examples of such problems include fraud detection, rare disease diagnosing, etc. However, the most commonly used classification algorithms do not work well for such problems because they aim to minimize the overall error rate, rather than paying special attention to the minor class. In the master's thesis, a number of models are evaluated with the objective to find those that better address the classification problem of imbalanced datasets. A special focus is given to the investigation of some ways of dealing with imbalanced datasets based on a Logistic Regression.