The prediction model uses a mathematical textual analysis that registers users' vocabulary

Predicting student's outcomes is at the heart of the most educational institutions. Parents are no different when it comes to knowing about the future their children will pursue. Artificial Intelligence (AI) and its applications have helped education sector figure out a lot of study related outcomes that will help students follow the right dream.

Artificial intelligence models are considered to be outstanding when doing predictive analysis. The technology anticipates the future by analyzing past data. One such extraordinary prediction was by an AI model used by Facebook to help find users when they get suicidal thoughts based on their feed posts. Other than that, artificial intelligence is highly utilized by the education sector to better inform students on where they should concentrate. However, since the outbreak of Covid-19, students have started getting education at home. Virtual classrooms and online tests are the most trending topics in the technology and education world.

In the middle of all this, a team of Russian researchers has used an AI-based model to predict high academic achievers from lower ones based on their social media posts.

AI model predicts student's educational outcomes

Ivan Smirnov, a leading research fellow of the laboratory of computational social sciences at the Institute of Education of HSE University, has created a computer model that can distinguish high academic achievers from lower ones based on their social media posts. The study was supported by a grant from the Russian Science Foundation (RSF), and an article detailing the study's results was published in EPJ Data Science.

The prediction model uses a mathematical textual analysis that registers users' vocabulary (its range and semantic fields from which concepts are taken), characters and symbols, post length, and word length. Every word is analyzed and rated based on its delivery. It is similar to finding the IQ of a student.

Students were ranked on the content they post on social media. For example, scientific and cultural topics, English words, and posts longer in length are ranked highly. These words represent good academic performance. However, an abundance of emojis, words or whole phrases written in capital letters and vocabulary related to horoscopes, driving and military services stipulate lower grades in school. At the same time, posts can be quite short but informative. This is taken at a high rank.

Smirnov's study used a representative sample of data from HSE University's longitudinal cohort panel study, Educational and Career Trajectories (TrEC). The study traces the career paths of 4,400 students in 42 Russian regions from high schools participating in PISA (the Programme for International Students Assessment). The study data also includes data about the students' VK accounts (a Russian social media and social networking service).

This kind of data, in combination with digital traces, is difficult to obtain and is seldom used. However, the dataset allows the researcher to develop a reliable model that can be applied to other settings. And the result can be extrapolated to all other students, both high school and middle school students.

The mechanism behind AI model

Posts from publicly viewable VK pages were used as a training sample including a total of 130,575 posts from 2,468 subjects who took the PISA test in 2012. The test allowed the researcher to assess a student's academic aptitude as well as their ability to apply their knowledge in practice.

In the study, unsupervised machine learning with word vector representations was performed on VK post corpus (totalling 1.9 billion words, with 2.5 million unique words). It was combined with a simpler supervised machine learning model that was trained in individual positions and taught to predict PISA scores.