A NEW PARADIGM FOR AUTISM DIAGNOSTICS: INTEGRATING MACHINE LEARNING AND VERBAL DECISION ANALYSIS

Autism Spectrum Disorder (ASD) affects millions of people globally, with

Early diagnosis is essential for effective interventions. However, the diagnostic process

Traditional is time-consuming and complex. This study proposes an innovative hybrid model to optimize

the diagnosis of ASD in children aged 0 to 5 years, combining Machine Learning (MA)

with Verbal Decision Analysis (ADL). The model uses the Random Forest algorithm to identify

the most relevant characteristics in the Medical Evaluation of the BPC/LOAS of the INSS and the

ZAPROS-III-i to sort them by importance. The results show a reduction of 89%

in the characteristics to be evaluated, speeding up the diagnosis without significant loss of accuracy.

The hybrid model, based on data from thousands of crianc.as diagnosed with ASD, presents

a new paradigm for improving and optimizing the diagnosis of ASD