The system predicts the risk of giving up online students

The system predicts the risk of giving up online students

Am Sure you can learn successful Blogging secrets through me, am also sure you can learn how to make money online with the help of my updates why not leave your email behind let me show you how.

Algorithms take into account up to 120 variables in creating profiles for each student, making night-time activity more associated with lower risk In distance learning, the dropout problem is traditionally higher, but the amount of potentially useful information is higher.

Students they can easily enroll on online university courses, but their completion is much harder. The rate for online courses can reach 80%. Researchers are trying to change this by developing early warning systems that suggest students likely to leave school. Administrators can then use forecasts to address more risky students and make efforts to keep these young people.

The largest system for such an early warning was developed in Spain – is based on the data of more than 11,000 students enrolled in online programs at UDIMA for five years.

The system is called the SPA (an abbreviation of Spanish, which means a system that prevents dropping). It uses a self-learning machine to generate individual estimates for new and permanent students, writes IEEE Transactions on Learning Technologies in an article published on April 16.

The ambitious university project, Juan Alcolea, Dimetrical’s Analytic Director, saw an interesting opportunity to have such a large set of data and attempted to associate with researchers at the online university as well as those at Universidad Autónoma de Madrid to develop a SPA through self-styling techniques. “It was clear that this is a problem that can benefit from new techniques,” Alcolea says.

“Particularly in distance learning where the abandonment problem is traditionally higher – but also the amount of potentially useful information”. SPA does not include only personal information (eg Age or Gender), economic data (e.g., type of payment), administrative data, academic estimates, and early/late registration data. It is also crucial to include behavior data from the online education system of the university.

This includes information such as time and duration of student activity. SPA can take up to 120 variables when designing profiles for each student. Then the total probability of falling is generated (for example, student A has a 60% chance of leaving school). Newer students have less data. This also determines the fact that the system is less accurate.

However, since self-learning machine learning algorithms take additional variables for students who are studying for many years, interesting models still appear. Algorithm Capture Trends Computer algorithms noticed interesting trends.

For example, age is an important factor in predicting the risk of abandoning new students. Those younger than 20 years of age give up more than older students. The way students share their daily learning activities is also very indicative – for example, whether they have studied especially during lunch or evening classes. Increased activity at noon is associated with a higher risk of bouncing while higher night-time activity is associated with lower risk.

The SPA also points out the gap between the sexes, where women are at greater risk of leaving school than men. “Indicators such as the amount, length and time between the pupils – between them and their teachers – or different characteristics based on activity trends – as an activity that remains constant, increases or decreases – seem to have low or no predictive power.

Contrary to what we initially thought – says Alcolea. Symptoms rather than problems However, such promising results must be unanimous, warns Susan Terio, lead researcher of the American Institute for Research, specializing in the development of Early Warning System for 12th grade Schools Terio is aware that Early Warning Systems, such as on-line SPA-based programs, may involve much more data than traditionally covered by the Early Warning System up to grade 12, where largely the absence/presence and estimates.

But, she thought, it was early to draw conclusions about the models that were ot waved modeling tools. “One of the things that is pretty clear is that analyzes show symptoms, not problems.” You can not diagnose only with information about the symptoms. Normally, you need to deepen deeper.

“Spanish specialists are also aware of this, suggesting that when online administrators of the curriculum identify students at risk of abandonment, they can contact students via e-mail or phone to talk to them in order to find a problem in each case. using the program in this way, says Alcolea, the research team is now planning to analyze the effectiveness of various detention measures.

The system predicts the risk of giving up online students

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *