Evaluation Metrics for Assessing the Performance of Diabetes Prediction Models

EasyChair Preprint no. 13593

Abstract

Evaluation metrics play a crucial role in assessing the performance of diabetes prediction models. These models aim to predict the likelihood of an individual developing diabetes based on various factors such as age, weight, family history, and blood glucose levels. Accurate evaluation of these models is essential to ensure their effectiveness and reliability. This paper provides an overview of commonly used evaluation metrics for assessing the performance of diabetes prediction models.

The evaluation metrics discussed in this paper include accuracy, sensitivity, specificity, precision, receiver operating characteristic (ROC) curve, area under the ROC curve (AUC), F1 score, and Matthews correlation coefficient (MCC). Each metric is defined, and its calculation method, interpretation, and limitations are explained. The paper emphasizes the importance of considering the goals and application of the model, as well as the trade-offs between different metrics, in order to choose the most appropriate evaluation approach.

Keyphrases: Diabetes Prediction, evaluation metrics, machine learning, model, Precision, sensitivity, specificity

@Booklet{EasyChair:13593,
  author = {Ayuns Luz},
  title = {Evaluation Metrics for Assessing the Performance of Diabetes Prediction Models},
  howpublished = {EasyChair Preprint no. 13593},

  year = {EasyChair, 2024}}