Company Announcement - Rumbi Maisva, a NMMU BCom (Hons) Information Systems student, has developed a system to bring knowledge closer to learners more quickly. The system automates the process of answering questions; hence assisting an educational mobile content site with its backlog of about 10 500 questions. The cost of accessing information is a major stumbling block especially for individuals who cannot afford the costs of full mobile Internet access using modern smart phones.
MXit, as a social media platform, provides a cost-effective alternative to the market leaders and content providers among growing communities of young learners desperate to access online content via older mobile phones (feature phones). An educational mobile content site accessed using MXit is Speak to Prof Ron, where learners from over 80 different countries submit questions relating to science, technology, engineering, maths and innovation through the Every1Mobile portal. Questions range from “why is the sky blue?” to “I want to become a doctor but I am doing Mathematics Literacy”. The concept was developed by Dr Ron Beyers of the Young Engineers and Scientists of Africa (YESA – yesa.org.za) in partnership with the Meraka Institute.
To date over 50 000 questions have been submitted in just 18 months and the demand for answers to questions is growing daily because it is currently a manual response system. Rumbi’s new automated system supports automated responses to learners’ questions where similar questions were previously answered and reside in the question database. The system analyses the questions even where there are spelling mistakes, grammatical errors and identifies existing questions in the database that are the best match.
“If the question is a new question for which there are no matches, it is routed to a human administrator to answer and subsequently becomes part of the question database. The others are matched up. In this way many more learners can be assisted immediately and automatically,” says the Computing Sciences student.