umap2016ea - analyzing mooc entries of professionals on linkedin for user modeling and personalized...
TRANSCRIPT
Analyzing MOOC Entries of Professionalson LinkedIn for User Modeling andPersonalized MOOC RecommendationsGuangyuan Piao, John G. BreslinInsight Centre for Data Analytics, National University of Ireland, Galway
Centre forData Analytics
Introduction
Massive Open Online Courses (MOOCs) play a significant role in educat-ing professionals. According to a recent study, over half of MOOC learners(62.4%) reported themselves as being employed full-time or self-employed.
Figure 1. LinkedIn functionality of adding finished MOOCs to user profiles.
Aim of Work
To investigate whether information in different fields of professionals’ profilesfrom LinkedIn allows to produce useful user profiles which can be used forpersonalized MOOC recommendations.
Three main fields of LinkedIn profile• job titles: Software Engineer, Java Engineer
• education fields: Information Engineering
• skills: Java, C++, Microsoft Excel
User Modeling Strategies
Figure 2. Skill-based user profiles.
Figure 3. Job-based user profiles.
Figure 4. Edu-based user profiles.
Experiment Setup
MOOC recommendations• dataset: 4,401 LinkedIn profiles (321 users for test set)• task: recommending MOOCs based on user profiles• recommendation algorithm:
sim (u, c) =~Pu
|| ~Pu||· ~Pi (1)
Evaluation Metrics• MRR: Mean Reciprocal Rank• S@N: Success rate at rank N
Results
0.19
0.23
0.21
0.29
0.24
0.33
0.26
0.35
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
MRR
S@05
performanceofrecommenda,ons
approa
ches
skill-based job-based edu-based pop
Figure 5. The quality of recommendations using different user modelingstrategies. pop denotes a non-personalized recommender which recommends
the most popular MOOCs among learners.The Insight Centre for Data Analytics is supported by Science Foundation Ireland under Grant Number SFI/12/RC/2289