Predictive risk-based system for predicting attention deficit hyperactivity disorder (ADHD)

DE202026105556U1Undetermined Publication Date: 2026-11-26HOWLADAR PAMPA +4
0 Cites 0 Cited by

Patent Information

Application Number
DE202026105556
Authority / Receiving Office
DE · DE
Patent Type
Utility models
Current Assignee / Owner
Filing Date
2026-09-09
Publication Date
2026-11-26
Estimated Expiration
2036-09-30
Patent Text Reader

Abstract

Predictive risk-based system for predicting attention deficit hyperactivity disorder (ADHD) in children, comprising: a data collection module configured to receive and store a high-dimensional survey dataset containing child health records that include demographic, socioeconomic, clinical, behavioral, environmental, and family information variables, while excluding post-diagnostic variables, treatment variables, severity indicators, behavioral outcomes, duplicate variables, age-derived variables, and sparse features from the dataset;a data preprocessing module configured to construct a binary target variable from the survey dataset, remove ambiguous labels, handle missing values ​​through a hybrid imputation approach, and address class imbalance to ultimately prepare the dataset for predictive modeling;a feature selection module configured to identify and retain predictive risk factors as predictor variables, wherein the feature selection module is configured to implement an effective embedding technique for identifying relevant features during model training, wherein the effective embedding technique enables the estimation of feature importance using measures such as mean decrease in impurity (MDI) and mean decrease in accuracy (MDA), thereby enabling the trained predictive model to capture complex nonlinear relationships, process mixed data types, and remain robust against noise and overfitting;a prediction engine configured to predict children with ADHD, using the final dataset prepared by the trait selection module, and training multiple machine learning classifiers using this final, balanced dataset, wherein the prediction engine classifies whether the child has ADHD or not based on the provided input, which includes predictive risk factors associated with ADHD; and a model interpretability module configured to quantify the contributions of individual risk factors to ADHD prediction and to visualize the dataset.
Need to check novelty before this filing date? Find Prior Art