Postal Savings Bank of China Deploys AI-Driven Big Data Project with Guidance of Central Bank

1431

One of China’s big six state-owned banks is making use of artificial intelligence (AI) and big data technology to augment operations at its Fujian province head branch.

Postal Savings Bank of China (PSBC) has launched an AI-driven “big data applications modelling project” at its Fujian province branch in the provincial capital of Fuzhou with the guidance of the Chinese central bank.

According to PSBC the project involves the operation of a unified big data modelling platform to apply machine learning technology to the bank’s targeted sales and risk control operations, as part of efforts to “fully unearth the value of big data and raise the digitisation level of financial services.”

“On the one hand the use of artificial intelligence to create a big data modelling platform provides comprehensive support to the branch’s shift from ‘specialist experience’ to ‘machine learning’ when it comes to data analysis work.

“On the other hand, the application of an AI big data modelling platform to client analysis, targeted sales and after-loan management operations fully uncovers the value of big data and raises the smart level of data analysis.”

The platform encompasses the six major functional modules of data management, feature processing, model generation, model applications, operation monitoring and visualised development, integrating model development, deployment and application and covering all stages of AI application.

“The platform provides graphic model development tools, adopts advanced machine learning algorithms and self-learning mechanisms, and possesses highly efficient distributed processing capability.”

PSBC has already used machine learning to develop credit card risk prediction models to forecast the probability that credit card users will be late in making payments, which it says that raised the accuracy and timeliness of risk management work.

According to PSBC the platform has reduced the “technological threshold” for operations modelling and reduced the time and labour required for data analysis projects.

“Since the platform came online, only one operations specialist and one data analyst are needed to complete an operations modelling development application within a period of as little as one to two months, cutting the time and labour needed in half compared to traditional methods and greatly reducing operating costs.”

PSBC said that machine learning models have also raised the accuracy of client sales targeting and risk prevention, helping it to strengthen its financial inclusion service offerings.