TOMSK, Jul 22 – RIA Tomsk. Ekaterina
Atamasova, a graduate student of TSU, has developed a methodology
allowing to create a psychological portrait of a person on his
"footprint" in social networks; a software prototype has already been
created to help build individual learning paths and improve the quality
of education, the university's press service said on Thursday.
it was reported that Tomsk State University (TSU) has been studying the
phenomenon of social networks in education since 2010 and put the
results into practice. Thus, employees of the university based on the
analysis of digital traces of high school students predict what profile
of training may choose an applicant, as well as train colleagues from
other universities to look for future students on the digital footprint.
Atamasova, a graduate student of the Institute of Applied Mathematics
and Computer Science of Tomsk State University, has developed a method
for selecting psychological characteristics in order to predict them
using a person's digital footprint. The prototype of the software, which
makes a psychological portrait based on open data obtained from social
networks, was created on its basis", – the report says.
is explained that to make a psychological profile is usually used
questionnaire or survey, but not everyone can take part in them.
Research of digital footprints is practically unlimited. Usually,
profile information, interests, friendships, texts and user activity are
used for analysis; some more information can be obtained from photo and
Atamasova identified 12 indicators of
user activity and proposed a methodology for selecting psychological
characteristics beyond the "big five. "To test the methodology, first of
all, it was necessary to collect data, namely the results of
psychological testing and digital footprint of respondents on the social
network VKontakte... On their basis for the computer model were
formed training samples", – she is quoted.
and employees of TSU with the help of LMS-system Moodle passed online
survey, the results of which revealed their psychological
characteristics. Then the computer model was "taught" to analyze the
desired characteristics by assessing the degree of their expression. The
skills acquired by AI were tested on the new user profiles and
conducted a survey to evaluate the accuracy of the computer model.
was found that stress can be predicted more accurately than
benevolence, which means that behavioral disorders can be identified and
individualized educational trajectories can be built, which will
improve the quality of learning. It is planned that the new software
will be introduced as a service for tutors and school psychologists.