This project formed the artefact for my MSc Thesis. From my personal experiences as a trainee teacher workload presents a major cause for otherwise suitable teachers dropping out of the profession.
Therefore, this project proposed the application of e-learning research and commercial systems into Adaptive Learning Systems in the context of learning resources for teacher's everyday planning. Integrating the older technology of using hypertext paths, such as that implemented by the Walden's Learning Paths project, into an ALS enabled the tagging according to an e-learning metadata standard of resources across the web without intruding upon the sources ownership of the content itself.
Objectives and Background:
The main objectives of the project were to:
- Implement Elasticsearch to enable flexible search according to users needs.
- Enable users to construct simple or complex learning paths.
- To ensure users are presented with key metadata about learning by implementing properties from the Learning Resource Metadata Initiative for the resource entity types in the application and search index.
- Retrieval of learning resources from across the web with metadata and displays varied and enhancing access to non-technical users of linked data sources of to consistently tag resources across the web.
- To provide a simple recommender to suggest resources according to learning theories by using relevant taxonomies stored as metadata and the Drupal views.
What I achieved and Skills Gained:
The main skills included:
- Installation and basic configuration of the Nutch web crawler and MongoDB.
- Understanding of the technologies required in order to integrate multiple services into an application and multiple applications in a monolithic architecture.
- Use of REST and curl requests for JSON data in Elasticsearch and MongoDB
- A knowledge of the variation of query and filter types available and stemming to tailor a search.
- Awareness of issues in configuring a search server with Drupal which impact the functionality of an application (e.g. entity reference view widgets addition links was not compatible with search API views at the time of this project.
Future additions to this project:
- Using Docker to test the system's integration using a Solr search server instead of Elasticsearch.
- Implemenetation of synonyms, stemming, fuzzy match terms and other search features of text analysis to tune the search server.
- Development of an API to integrate with Learning Management Systems such as by using the Learning Tools Integration API.
- The development of an ontology to provide a conceptual learning recommender.
Recommendations for similar future projects
- To implement improved recommender algorithms using machine learning algorithms which provide flexible recommendations according to teaching theories.
- To design a web crawler which uses machine learning to tag resources with metadata according to the standards reasonably accurately