Handling varying amounts of missing data when classifying mental-health risk levels

One of the main challenges of classifying clinical data is determining how to handle missing features. Most research favours imputing of missing values or neglecting records that include missing data, both of which can degrade accuracy when missing values exceed a certain level. In this research we propose a methodology to handle data sets with a large percentage of missing values and with high variability in which particular data are missing. Feature selection is effected by picking variables sequentially in order of maximum correlation with the dependent variable and minimum correlation with variables already selected. Classification models are generated individually for each test case based on its particular feature set and the matching data values available in the training population. The method was applied to real patients' anonymous mental-health data where the task was to predict the suicide risk judgement clinicians would give for each patient's data, with eleven possible outcome classes: zero to ten, representing no risk to maximum risk. The results compare favourably with alternative methods and have the advantage of ensuring explanations of risk are based only on the data given, not imputed data. This is important for clinical decision support systems using human expertise for modelling and explaining predictions.

Knowledge management in support of enterprise risk management

Risk management and knowledge management have so far been studied almost independently. The evolution of risk management to the holistic view of Enterprise Risk Management requires the destruction of barriers between organizational silos and the exchange and application of knowledge from different risk management areas. However, knowledge management has received little or no attention in risk management. This paper examines possible relationships between knowledge management constructs related to knowledge sharing, and two risk management concepts: perceived quality of risk control and perceived value of enterprise risk management. From a literature review, relationships with eight knowledge management variables covering people, process and technology aspects were hypothesised. A survey was administered to risk management employees in financial institutions. The results showed that the perceived quality of risk control is significantly associated with four knowledge management variables: perceived quality of risk knowledge sharing, perceived quality of communication among people, web channel functionality, and risk management information system functionality. However, the relationships of the knowledge management variables to the perceived value of enterprise risk management are not significant. We conclude that better knowledge management is associated with better risk control, but that more effort needs to be made to break down organizational silos in order to support true Enterprise Risk Management.