This paper proposes a data-driven solution to assess job skill value from a market-oriented perspective. The task is formulated as a Salary-Skill Value Composition Problem, where job salary is influenced by the context-aware value of required skills. A cooperative neural network, Salary-Skill Composition Network (SSCN), is proposed to separate and measure skill value from job postings. Experiments show SSCN effectively assigns skill values and outperforms benchmark models for salary prediction.
