Abstract: Author name disambiguation is an important basic problem in scientific literature research, which has not been well solved in Chinese names. This study aims at improving the accuracy rate of Chinese name disambiguation. This paper proposes an algorithm based on author's career experience and citation networks for the first time. The F1-score of this algorithm conducted on the ground truth dataset of Chinese authors' papers from Web of Science reaches to 80.88%. This algorithm has some limitations in data availability and large-scale use. The algorithm proposed in this paper is the first time to eliminate name disambiguation among WOS Chinese authors. It has the characteristics of strong operability, fast operation speed, independent of complex models and not limited by computing resources, and has a good application prospect. The ground truth dataset constructed in this paper also has important reference significance for the follow-up research.

Key words: Name Disambiguation, Chinese Names, Supervised Learning, Career Experience, Citation Networks