American researchers have asked to depict children engaging in things and people. They tried Eight models of automatic learning "This helps you to control the procedure and to make it more efficient," according to a study published by the scientific journal PSOS Medicine.
The study was conducted in a study from Stanford University School of Medicine. At that time, Dennis Wall, professor of Pediatrics and Biomedical Data Science, headed the city of California.
Each model contains a set of algorithms There were also characteristics of children 5 to 12 A general score generated by the child indicates whether autism is involved, "he explained.
How the videos are handled
Wall learns that the evaluation of the models is evaluated. They wanted home-based videos from one to five minutes long recruited for study. Displaying children's faces and hands and using "their social interaction, toys, pencils and utensils".. From these pictures, 116 boys with an average age of 4 years and 10 months have been found to have a history of autism. An average of 46 children had been developing it for two years and 11 months.
One of the expert specialists analyzed a video using videos Question 30 Autism was later included into eight mathematical models by answering the "yes" or "no" based on the nature of nature.
Outstanding children with sampling autism are 94.5% of children with autism and 77.4% of children without autism. They evaluate the result to evaluate The other 66 videosHalf of autism children. The same model has a child with 87.8% autism and 72.7% of the disease.
Another advantage of using home videos Diagnosis "They accept the kid in their natural environment", unlike the clinical assessment in a medium, "making a hard and artificial divergent behavior". "We've proven that we can identify small groups with differential qualities." In professional consultations, experts can quickly and independently evaluate within a virtual online environment within minutes. "