What influence digital inclusive finance from policy tones? Based on machine learning

Tianyi Lei, Jingjing Jiang,
What influence digital inclusive finance from policy tones? Based on machine learning,
Finance Research Open,
Volume 1, Issue 4,
2025,
100058,
ISSN 3050-7006,
https://doi.org/10.1016/j.finr.2025.100058.
(https://www.sciencedirect.com/science/article/pii/S3050700625000581)
Abstract: This paper investigates the relationship between regional policy tones and the development of digital inclusive finance (DIF) in China. First, we employ the textual analysis method to collect policy text data from 335 prefecture-level cities in China, and then construct a policy tone matrix. Next, we use machine learning methods, including BP neural networks, LSTM, and decision trees, to simulate the actual relationship between policy tones and DIF, thereby clarifying the complex interplay between the two. The results show that the environmental policies of the notification and advisory tones significantly contribute to the development of DIF. Specifically, the advisory environmental policies have a substantial impact on the coverage breadth of DIF. Meanwhile, the economic policies, including notification, regulatory, and advisory tones, significantly promote the development of DIF. This study contributes to the existing literature on the factors influencing DIF and provides valuable insights for practical policy-making.
Keywords: Policy tone; Digital inclusive finance; Textual analysis; Machine learning