Automatic creation of e-learning content using blog posts

Abstract

In the contemporary world, blogging is no longer used just for personal stuff and providing expert opinions, blogs often carry useful “learning” content, but most of the time in unstructured manner. Tech companies often use internal and external blogs to allow their employees to share the expert knowledge within the company instead of incorporating this knowledge with existing e-learning tools like virtual learning environments as it takes an extra effort and cost to integrate the knowledge in the blogs with the e-learning tools. Deriving elearning course content from blogs helps those companies as they can still use their own blogs and adopt e-learning tools without much burden and cost. This study presents a model to automate the process of integrating the new knowledge from the blogs with an e-learning platform. There are three main issues addressed in the proposed model: identifying the appropriate methods to extract knowledge content from blog posts, extract the content from blogs using identified methods and create learning content, integrate the created learning content with selected e-learning tool. Finally, the developed model was run against the selected blog posts and the created e-Learning content was evaluated using human judgement. Normally a blog page contains boilerplate and a lot of clutter like pop-ups, extra links and unnecessary icons. Hence extracting the “meat” of a blog page and eliminating the clutter is an eminent task. Content extraction from webpages, blogs is heavily studied area as there are lot of applications like sentiment analysis, content summarization and text classification. Also, there are already built open source and proprietary tools to extract the core content from the web pages. Each of them has their own merits. This study evaluates python based three content extractor libraries: Readability, Boilerpipe and NewsPaper to find out which one is giving the best results. This study revealed that among six of the selected extractors, both Readability and NewsPaper libraries outperform other four extractors in Boilerpipe library. When comparing Readability and NewsPaper libraries, Readability library has a slight edge over the NewsPaper library as it can extract code segments and emojis as well. When integrating the extracted content with the iii Moodle, an already available Forum has been used. The content was transferred as a new discussion under the forum using Moodle REST API functions. The proposed model can extract the content from a blog and then extracted content transferred to Moodle. Without manually creating the course content by copy and pasting, trainers and technical experts can easily integrate their blogs with an e-Learning system.