ADHD & Gender Prediction

Interpretable clinical ML

2025

Built multi-output classification models to predict ADHD diagnosis and participant gender from behavioral assessment data.

Key achievements

  • Benchmarked Logistic Regression, kNN, and XGBoost with feature engineering and evaluation techniques.
  • Applied SHAP-style explanations to make model outputs interpretable for clinical researchers.

Project narrative

Analysed ADHD assessment data to identify behavioural patterns across male and female participants. Built multi-output classification models (Logistic Regression, kNN, XGBoost) to predict both ADHD diagnosis and participant gender.

Applied feature engineering, model tuning, and evaluation techniques to improve model explainability and performance. SHAP-style attributions were translated into clinician-friendly narratives for individualized interventions.

The project contributed to research on gender-based differences in ADHD for more targeted assessment approaches, achieving a 0.82 F1 score while improving sensitivity for underrepresented cohorts by 14%.

Key words

Filter-ready keywords covering tech stack, methodologies, and differentiators.

Healthcare AIFeature EngineeringExplainabilityLogistic RegressionXGBoostGender Insights

Relevant projects

Projects that share similar methodologies, tech stacks, or impact areas.

Gesture Prediction

sEMG-driven hand gesture classification

2025

Analyzed time series sEMG data and developed deep learning and ensemble models to predict 10 hand gestures.

  • Extracted 96 time- and frequency-domain features per segment from raw biosignals.
  • Achieved 82% multi-class average accuracy using XGBoost.
Signal ProcessingTime-Series MLFeature Engineering
Download
View project

Data Viz Pro

No-code analytics studio

2025

Built a Streamlit platform that guides beginners from raw datasets to trained models with side-by-side insight panels.

  • Launched a guided workflow covering upload, cleaning, visualization, modeling, and export.
  • Implemented dual-panel previews so users see how preprocessing changes affect plots and tables.
Data VisualizationStreamlitModel Ops
View project

Galaxy Type Classification

Multi-class astronomical classification

2024

Developed predictive models for multi-class galaxy classification using regression techniques.

  • Achieved 88.78% accuracy with a multinomial logistic regression model.
  • Manually built multinomial regression from individual logistic regression base models to validate the approach.
RegressionVisualizationAnalytics
View project