Article: Knowledge graph-enhanced reinforcement learning for generating students personalised career progression sequences Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.80 pp.26 - 50 Abstract: Students often fail to align course selection with long-term career goals due to overwhelming resources and lack of career-oriented guidance. Existing learning path methods emphasise short-term performance but neglect vocational skill accumulation, resulting in suboptimal career preparation. This paper proposes a knowledge graph-enhanced reinforcement learning framework generating personalised course sequences for career advancement. Grounded in social cognitive career theory, it integrates self-efficacy, outcome expectations, and personal goals into state and reward mechanisms. A knowledge graph integrating courses, skills, and occupations from occupational information network is embedded via a graph convolutional network to provide skill-aware student representations. A deep Q-network optimises sequences by maximising a dual-objective reward balancing grades and career alignment. Experiments show this framework achieves a 68.5% career achievement rate, outperforming the strongest baseline by 14.2 percentage points, while reducing average courses to reach a career goal to 5.6. Inderscience Publishers - linking academia, business and industry through research

Title: Knowledge graph-enhanced reinforcement learning for generating students personalised career progression sequences

Authors: Miao Chu

Addresses: School of Economics and Management, Wuchang Institute of Technology, Wuhan, 430065, China

Abstract: Students often fail to align course selection with long-term career goals due to overwhelming resources and lack of career-oriented guidance. Existing learning path methods emphasise short-term performance but neglect vocational skill accumulation, resulting in suboptimal career preparation. This paper proposes a knowledge graph-enhanced reinforcement learning framework generating personalised course sequences for career advancement. Grounded in social cognitive career theory, it integrates self-efficacy, outcome expectations, and personal goals into state and reward mechanisms. A knowledge graph integrating courses, skills, and occupations from occupational information network is embedded via a graph convolutional network to provide skill-aware student representations. A deep Q-network optimises sequences by maximising a dual-objective reward balancing grades and career alignment. Experiments show this framework achieves a 68.5% career achievement rate, outperforming the strongest baseline by 14.2 percentage points, while reducing average courses to reach a career goal to 5.6.

Keywords: knowledge graph; reinforcement learning; personalised learning path; career development; social cognitive career theory.

DOI: 10.1504/IJICT.2026.155106

International Journal of Information and Communication Technology, 2026 Vol.27 No.80, pp.26 - 50

Received: 20 Mar 2026
Accepted: 18 May 2026
Published online: 27 Jul 2026 *

Title: Knowledge graph-enhanced reinforcement learning for generating students personalised career progression sequences

Keep up-to-date