The Learning Object Relation Discovery (LORD) plugin is now available from the Plugins directory (https://moodle.org/plugins/block_lord). This plugin extracts the text content from the learning activities in a course and uses that content to compare the activities against each other, producing a similarity score between each pair of activities. The comparison uses an online sentence similarity calculator (https://ws-nlp.vipresearch.ca/) to determine the similarity values. The similarity calculator uses WordNet, so is currently limited to the English language, but support for Hindi and French are planned for later in 2021. The comparison process runs as a background task and can take some time to complete.
Once all the similarity values are known between all the learning objects, the learning object graph (LOG) can be made. A network graph is created where the nodes are the learning objects from the course and the similarity values between objects is used to position the nodes, grouping learning objects with similar content.
This plugin can be used by itself to examine the similarity between learning activities and also integrates with the Behaviour Analytics plugin (https://moodle.org/plugins/block_behaviour). The Behaviour Analytics plugin uses a network graph of learning objects to cluster students according to their behaviour patterns, but the graph nodes are grouped by section and require manual arrangement. The LORD plugin is designed to automatically configure the graph so that it does not have to be manually configured, which should produce better clustering.