Computer algorithms that help children
PhDr. Ing. Marek Dobeš, PhD. from the Institute of Social Sciences CSPS SAS and doc. Ing. Peter Drotár, PhD. from the Faculty of Electrical Engineering and Informatics of the Technical University of Košice are currently concerned with the question of to what extent is it possible for a computer algorithm to reveal from a font sample whether a child may be dysgraphic. This research is carried out within the grants of APVV "Computer decision support system for hepatic encephalopathy" and VEGA "Computer model of integration of chemosensory modules and motor module of the neural network C. Elegans".
"Dysgraphia is a writing disorder when a person writes illegibly, mistakes the letters, or his writing is otherwise distorted. Because writing is a strong part of our culture and a lot of attention is paid to it in schools, children with dysgraphia may have problems in academic life. Moreover, these problems can be further reflected in their overall satisfaction at school as well as in private life,” said Marek Dobeš from the Institute of Social Sciences CSPS SAS.
That is why early diagnosis of dysgraphia is very important, according to the scientist. Nowadays, the diagnosis of dysgraphia is performed by professionals who have been trained in this field. However, this means that their number is limited and does not allow every child to be examined. In this way, it may happen that a child suffering from dysgraphia does not have to be diagnosed with it, and therefore may encounter problems in school practice. Dysgraphia can be mistaken for laziness in writing, or a little willingness to practice writing at home.
"Computer algorithms come into play here. Progress in computer science allows computers to process real-world data and classify it with similar, sometimes greater, accuracy than humans. They are used, for example, in the diagnosis of tumours from X-rays or in the analysis of magnetic resonance imaging. A lot of data is presented to the computer which tries to extract important features from them that distinguish the individual categories of data from each other,” said Marek Dobeš.
In dysgraphia screening, the child writes a few sentences on paper that is placed on a tablet adapted to it. The tablet saves not only the graphic expression itself. It monitors how long the child writes, what breaks he takes, how hard he presses on the pad, how high he lifts the pen and dozens of other data. It thus has an advantage over a professional examination by a person who usually analyses only a few characteristics of the font.
"Subsequently, the algorithm compares the values of these parameters with the values that were presented to it in the past. Within a few seconds, it shows the probability with which the child may have dysgraphia. If this probability is quite high, the child's parents may be advised to visit an expert. In this way, it is possible to examine many more children and enable them to treat this disorder in time,” added the scientist.
The research is in the testing stage of individual algorithms, and preliminary results show that their predicting power is around 80.
Image: unsplash/Annie Spratt