Utilizing Big Data Technology for Online Financial Risk Management

Abstract

The rise of cloud computing, internet of things and information technology has made big data technology a common concern for many professionals and researchers. A financial risk control model, known as the MSHDS-RS model, was creatively suggested in response to the present state of inappropriate feature data design in big data risk control technology. The concept is built on multi source heterogeneous data structure (MSHDS) and random subspace (RS). This model is novel in that it uses a normalized sparse model for feature fusion optimization to create integrated features after extracting the hard and soft features from loan customer information sources. Subsequently, a base classifier is trained on the feature subset acquired via probability sampling, and its output is combined and refined by the application of evidence reasoning principles. The accuracy improvement rate of the MSHDS-RS method is approximately 3.0% and 3.6% higher than that of the current PMB-RS methods under the conditions of soft feature indicators and integrated feature indicators, respectively, according to an observation of the operation results of MSHDS-RS models under various feature sets. As a result, the suggested optimization fusion approach is trustworthy and workable. This study has helped to reduce financial risks associated with the internet and may be useful in helping lenders make wise judgments.