FLAML-Boosted XGBoost Model for Autism Diagnosis: A Comprehensive Performance Evaluation
Dheiver Francisco Santos · Qeios · 2023
In this article, we address the challenge of imbalanced classification using automatic machine learning (AutoML) techniques in a case study on autism diagnosis. By leveraging the FLAML library, we demonstrate the process of balancing the dataset, training an XGBoost model, and evaluating its performance using various metrics, such as ROC curve, calibration curve, confusion matrix, and precision-recall curve. The XGBoost model achieves the best error of 0.0077, and the metrics provide a comprehensive view of its discriminative ability, calibration, and overall performance in autism classification.