Text Mining Experts in Bielefeld enhance Distance Education Internet Portal with Automatic Text Analytics

Internet portals that ask customers to assess products or services often face a major problem: how to analyze unstructured free texts and extract the relevant information. This is usually a very expensive and time-consuming process. A new research project launched by the CITEC research group “Semantic Computing” is addressing this problem in collaboration with the company “Online Akademie” (OAK). OAK runs an Internet portal called “fernstudiumCheck.de” that students can use to evaluate distance learning universities, classes, and programs.

FernstudiumCheck.de is an online platform on which people can submit or read reviews of distance education services and programs throughout Germany. To evaluate a course or program, users award one to five stars by categories such as the curriculum or the quality of supervision. “Such a review is easy to analyze statistically,” explains Dr. Roman Klinger, the head of the new research project. However, each reviewer is also asked to write a short report on her or his own personal experiences. “To analyze these free texts statistically, the content of every single sentence has to be sorted into content categories—and that is what makes the analysis so complicated and time-consuming,” says Klinger. “Up to now, OAK has been doing this painstakingly by hand. Our software will automatically mine the full content of these unstructured free texts,” says the computer scientist who is a member of the Semantic Computing research group at CITEC. The head of this research group, Professor Dr. Philipp Cimiano, adds, “we are making it possible for Online Akademie to give even more efficient feedback to the providers of distance learning programs so that they will have a detailed understanding of what needs to be improved and what they are already doing well.”

Roman Klinger and his colleagues are studying how people process natural language and how computers can imitate this processing. “We will exploit insights on how people process natural language within the project,” he explains. “To classify the content of these texts, we will rely on statistical procedures that can automatically estimate which category a sentence or phrase belongs to.” For example, the software will recognize whether the sentence “Most of the time, the teachers gave a quick response to my e-mails” should be assigned to the categories “supervision,” “course content,” or “cost-effectiveness.” The software will also determine whether the evaluation is positive or negative. “We hope that this will enable us to provide the distance learning universities with even better and more specific feedback for their quality management,” says Thomas Tibroni, the managing director of OAK.

The new project aims to make analyses faster and more precise. “In the future, the site provider should only need to press one button to get the complete analysis—and should be able to do this at any time,” says Klinger. The name of the project is Sentiment Analysis for Distance Education Evaluation (SADE).

The Semantic Computing Group is working on methods that will enable people to use tools to access large-scale knowledge resources in an intuitive way. The head of the group is Professor Dr. Philipp Cimiano. To simplify access to large bodies of knowledge, he and his team are developing ways to analyze the meaning of texts automatically.

The Online Akademie GmbH & Co. KG operates a range of portals dealing with topics in education, university studies, distance learning, and further education. Alongside the FernstudiumCheck portal, it also runs the evaluation portal StudyCheck.de.

Further information is available online at:
www.sc.cit-ec.uni-bielefeld.de

Contact:
Dr. Roman Klinger, Universität Bielefeld
CITEC research group "Semantic Computing"
Telephone: 0521 106- 12017
Email: rklinger@cit-ec.uni-bielefeld.de