Demans’ Patients’ Accident Predict AI System Prepared for Human Control

Although data bases related to traffic accidents are increasingly offering more automated security analysis, it is still a rather challenging task to convert artificial intelligence models into real decisions using this data. Especially to predict the severity of accidents mixed by demans patients, it becomes complex even if it is not limited to simple classification problem.

The researchers redefined this problem and addressed the topic as a “Selective melting frame”**, which will support the review of the experts of **insan. The tasks of the system were organized as follows:

• Breaks into three categories: no injury/only material damage, mild or moderate violence injury, death or severe injury
• Automatic decision-making in cases where the model is safe, waiting for man to decide if in uncertain circumstances
• Make predictable outputs to increase model calibration and auditability
• Designing system that can be integrated with work flows in the public sector

This approach reveals a perspective that artificial intelligence models only consider responsibilities and transparency in real life applications, beyond classification accuracy. The system is highlighted as a requirement to accurately identify high risk situations and present human decisions. Thus, both the effectiveness of technology and the end of man is preserved.