The impact of imbalanced training data on machine learning for author name disambiguation

TL;DR

This paper investigates how the ratio of negative to positive training data affects machine learning models for author name disambiguation, revealing that using less negative data can be nearly as effective and more efficient.

Contribution

It demonstrates that effective author name disambiguation models can be trained with less negative data, challenging the common practice of using all available data, and highlights the impact of data imbalance.

Findings

Increasing negative data yields marginal performance gains.

Logistic Regression and Naive Bayes perform well with balanced data.

Random Forest performance saturates after a 1:10 to 1:15 ratio.

Abstract

In supervised machine learning for author name disambiguation, negative training data are often dominantly larger than positive training data. This paper examines how the ratios of negative to positive training data can affect the performance of machine learning algorithms to disambiguate author names in bibliographic records. On multiple labeled datasets, three classifiers - Logistic Regression, Na\"ive Bayes, and Random Forest - are trained through representative features such as coauthor names, and title words extracted from the same training data but with various positive-negative training data ratios. Results show that increasing negative training data can improve disambiguation performance but with a few percent of performance gains and sometimes degrade it. Logistic Regression and Na\"ive Bayes learn optimal disambiguation models even with a base ratio (1:1) of positive and negative training data. Also, the performance improvement by Random Forest tends to quickly saturate roughly after 1:10 ~ 1:15. These findings imply that contrary to the common practice using all training data, name disambiguation algorithms can be trained using part of negative training data without degrading much disambiguation performance while increasing computational efficiency. This study calls for more attention from author name disambiguation scholars to methods for machine learning from imbalanced data.