Financial institutions put tremendous efforts on the data analytics work associated with the risk data in recent years. Their analytical reports are yet to be accepted by regulators in financial services industry till early 2019. In particular, the enhancement needs to meet the regulatory requirement the APRA CPG 235. To improve data quality, we assist in the data quality analytics by developing a machine learning model to identify current issues and predict future issues. This helps to remediate data as early as possible for the mitigation of risk of re-occurrence. The analytical dimensions are customer related risks (market, credit, operational & liquidity risks) and business segments (private, wholesale & retail banks). The model is implemented with multiple Long Short-Term Memory ("LSTM") Recurrent Neural Network ("RNNs") to find the best one for the quality & prediction analytics. They are evaluated by divergent algorithms and cross-validation techniques.
Wong, Ka Yee; Huang, Haojie; and Wong, Raymond, "Learning Data Quality Analytics for Financial Services" (2019). PACIS 2019 Proceedings. 67.