In recent decades, the indisputable change on the labour market led to a significant gap between skills demand and supply. This phenomenon creates major challenges for all stakeholders in education (Colombo et al., 2018; T Smith and Ali, 2014; Wowczko, 2015), but most prominently for lifelong learners, who need to monitor and update their individual skillsets on a regular basis to remain employable (Colombo et al., 2018). To facilitate this continuous need for re-skilling, learners need large amount of educational content tailored to their own needs. In this context, Open Educational Resources(OERs) have a high potential to supply learners with personalized learning content. However, the applicability of OERs in this area has been limited due to the lack of high-quality personalized services such as tailored recommendation and search (Chicaiza et al., 2017; Sun et al., 2018). In this research, therefore, we want to:
1. Define a method to help learners to set individual skill targets.
2. Collect valuable feedback from participants/attendees in order to
3. build a personalized OER recommender, providing learning content to master target skills.
First we constructed a model to decompose jobs into required skills, based on information received from job boards. Our model used sample vacancies from monster.com to identify skill-related sentences by applying word-level n-gram and multinomial logistic regression. Consequently, we applied TFIDF to detect skills in skill-related sentences. Finally, we counted the occurrences of skills and applied a simple decay function to calculate their importance in a particular job.
For our OER recommender system, we focussed on the following learner properties:
2.Target skill Levels
4.OER preferences (length, source, quality-assurance, accessibility)
To set the baseline for recommendations, learners need to address the first three properties. Based on these personal features, our recommender utilizes similar user profiles to initialize learner preferences. Similarly, for OERs, we use a number of properties (i.e. properties for subject, author, level, length, quality, accessibility) and benchmark those to similar OERs to initialize the missing properties.
During the learning process, learners rate their satisfaction after completing each recommended OER. Based on ratings, we update the properties of both users and OERs. This strategy helps us detecting the changes in user profiles, updating properties of OERs, finding missing properties of OERs, and ultimately, fine tune the precision of personalized content recommendations for individual skill targets.
Based on our design, we built a prototype dashboard, in which learners can search for their current or desired job, display the list of required skills, set their level of expertise for each skill, and for each target skill (ordered according to skill importance) get relevant OERs until the learner reports mastery in those skills.
This study is expected to 1, empower learners to take control and responsibility for their own development, and 2, improve their skills on the basis of labour market information and related OERs. We expect that learners will show enhanced self-regulation, thus spend less time and effort to find related, high-quality OERs. Furthermore, this approach will contribute to OER property identification efforts, which are essential to increase the levels of OER (re)usability both by instructors and learners.
Chicaiza, J., Piedra, N., Lopez-Vargas, J., Tovar-Caro, E., 2017. Recommendation of open educational resources. An approach based on linked open data, in: 2017 IEEE Global Engineering Education Conference (EDUCON)., IEEE, Athens, Greece, pp. 1316–1321. https://doi.org/10.1109/EDUCON.2017.7943018
Colombo, E., Mercorio, F., Mezzanzanica, M., 2018. Applying machine learning tools on web vacancies for labour market and skill analysis 31.
Sun, G., Cui, T., Xu, D., Shen, J., Chen, S., 2018. A Heuristic Approach for New-Item Cold Start Problem in Recommendation of Micro Open Education Resources, in: Intelligent Tutoring Systems. Springer International Publishing, Cham, pp. 212–222. https://doi.org/10.1007/978-3-319-91464-0_21
T Smith, D., Ali, A., 2014. Analyzing Computer Programming Job Trend Using Web Data Mining. Issues Informing Sci. Inf. Technol. 11, 203–214. https://doi.org/10.28945/1989
Wowczko, I., 2015. Skills and Vacancy Analysis with Data Mining Techniques. Informatics 2, 31–49. https://doi.org/10.3390/informatics2040031