Scholarlink

Research activity into the use of machine learning models for Autism-Spectrum-Disorder detection has experienced rising interest during the previous few years. The authors studied the ensemble model that connects XGBoost with LightGBM and CatBoost to improve ASD classification performance. The data preprocessing includes encoding variables and normalization along with ADASYN implementation for handling unbalanced classes. The model reaches its highest performance level through hyperparameter tuning along with Voting Classifier integration that contains beneficial aspects from all individual models. The SHAP analysis serves to make ASD classification more interpretable by identifying crucial characteristics which determine ASD outcomes. The experimental test results demonstrate that strong models work effectively after merging heterogeneous classification systems. The research uses an evaluation of multiple ML models and techniques to create a dependable automated detection system for ASD. The research findings contribute knowledge to AI-assisted healthcare because they enable scalable clinical solutions which enhance early identification and intervention practices. Improving autism spectrum disorder diagnosis through ML depends on selecting important features and adjusting class distributions without compromising the understanding of diagnostic outcomes.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Ensemble Learning for Autism Diagnosis: A High-Accuracy Machine Learning Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Manoj Masule,
  • Shailesh Galande,
  • Anushka Lahare,
  • Shreya Mohod,
  • Harshali Patil

摘要

Research activity into the use of machine learning models for Autism-Spectrum-Disorder detection has experienced rising interest during the previous few years. The authors studied the ensemble model that connects XGBoost with LightGBM and CatBoost to improve ASD classification performance. The data preprocessing includes encoding variables and normalization along with ADASYN implementation for handling unbalanced classes. The model reaches its highest performance level through hyperparameter tuning along with Voting Classifier integration that contains beneficial aspects from all individual models. The SHAP analysis serves to make ASD classification more interpretable by identifying crucial characteristics which determine ASD outcomes. The experimental test results demonstrate that strong models work effectively after merging heterogeneous classification systems. The research uses an evaluation of multiple ML models and techniques to create a dependable automated detection system for ASD. The research findings contribute knowledge to AI-assisted healthcare because they enable scalable clinical solutions which enhance early identification and intervention practices. Improving autism spectrum disorder diagnosis through ML depends on selecting important features and adjusting class distributions without compromising the understanding of diagnostic outcomes.