ML Hazard Auto Detection| Automated Detection of Hazards
Awards & Nominations
ML Hazard Auto Detection has received the following awards and nominations. Way to go!
Automated Detection of Hazards
AutoML Hazard Detection
Summary
An increase in the reliability of Health Information System(HIS) will facilitate institutional trust and credibility of the systems.To present an end-to-end framework for improving the reliability and performance of HIS. Specifically, describing the system model, present some of the methods that drive the model, and discuss an initial implementation of two of the proposed methods using data from the Veterans Affairs HIS and Corporate Data Warehouse systems. include (1) the design of a system model for monitoring and detecting hazards in HIS, (2) a data-driven approach for analysing the health care data warehouse.
How I Addressed This Challenge
**To develop an ML model for monitoring and detecting hazards in Health Information System, (2) a data-driven approach for analysing the health care data warehouse.
**this will help the health practitioner a quick response to their problem, easy detection.
**The model will detect and predict unforeseen hazards.
**By training the model will alot of data on different hazards scenarios.
**Using Neural Network(NN) Deep learning and Tension flow.
How I Developed This Project
The HMD component comprises of two layers: eventbased hazards detection and system-based reliability performance evaluation. This layer uses data-driven approaches to identify potential hazards for further investigation. The layer assumes that hazards are not known a priori. This assumption, thus, provides not only an opportunity to discover previously unknown or undetected hazards but also allows for analysts to remain in the loop. The analysts can review the detected hazards and determine whether they are true hazards or just noise in the data. These reviews will serve as feedback to the hazard-detection methods so that future occurrences are characterised correctly. This is a dynamic approach in which the detectors are continuously trained and improved using the analysts’ analogous to the concept of incremental learning in machine learning literature. Using this approach, we developed a series of hazard detectors (HDs); one HD called anomalous cancelled orders is discussed in this paper.
How I Used Space Agency Data in This Project
very interesting... i will to participate next year though this one reach me late.
Project Demo
The proposed system model comprises of three components that allowed for: (i) a multiple-stage hazarddetection framework and (ii) safety analysts involvement in the decision process. The iterative nature of the model allows for updating the logic behind the existing detectors and new insights to develop new detectors. Working with the VA’s VistA and CDW systems provides a platform for working with a largescale EHR databases. Furthermore, the anomalous cancelled order detector was shown to detect previously unknown hazards related to cancelled/ rejected orders. Similar methods could be used for datasets with similar characteristic features in HIT systems. The system surveillance model gives encouraging results in identifying critical state transition in a clinical workflow modelled as a MC; this could serve as a basis for describing potential failure scenarios or reliability degradation scenarios.
Data & Resources
Adelman, J. S., Kalkut, G. E., Schechter, C. B., Weiss, J. M., Berger, M. A., Reissman, S. H., . . . Southern, W. N. (2012). Understanding and preventing wrong-patient electronic orders: A randomized controlled trial. Journal of the American Medical Informatics Association, 20(2), 305–310. Bates, D. W., Cohen, M., Leape, L. L., Overhage, J. M., Shabot, M. M., & Sheridan, T. (2001). Reducing the frequency of errors in medicine using information technology. Journal of the American Medical Informatics Association, 8(4), 299–308. Bates, D. W., & Gawande, A. A. (2003). Improving safety with information technology. New England Journal of Medicine, 348(25), 2526–2534.