Expert assessments in decision making: risks and safety

The monograph “Expert assessments in decision making: risks and safety” presents studies that address the problems of expert assessments and the formation of decisions based on expert opinions in some areas of human activity. The results of decisions based on such assessments have a direct impact on safety and the risks that may follow. The problems are considered in the following aspects.

An innovative approach to the formation of a list of types of forensic examinations and expert specialties.
It has been established that the existing classification of forensic examinations by branches of specific examination used in their conduct is outdated and does not meet modern requirements. This creates difficulties not only for forensic experts and persons wishing to become them, but also for the court, participants in the trial and all citizens interested in obtaining a forensic expert's opinion. A discrepancy has been established between the types of forensic examination and the specialties of experts for which the qualification of a forensic expert is assigned in departmental lists.

It has been proposed to unify the interdepartmental approach to the classification of forensic examinations in order to avoid errors that could lead to illegal and unfounded court decisions. The issue of unification can be resolved by developing a general approach to the classification of types and types of forensic examinations, enshrined in an interdepartmental regulatory act, which should be based on the criteria of the general theory of forensic examination. The creation of a unified list of types of forensic examinations and their corresponding expert specialties can open up ways to solve many issues facing the expert community. Taking into account the fact that at the time it is planned to provide opinions by experts in electronic form, the creation of a forensic expert’s office and its integration with the Unified Judicial Information and Telecommunication System, the creation of a modern unified list by types of forensic examinations and their corresponding expert specialtiesis on the agenda.

Blood cell image recognition using texture and neural networks for leukemia diagnosis.
Morphological analysis of blood cell images is usually performed manually by an expert, but this method has many disadvantages, including slow analysis, low accuracy, and the results depend on the skill of the operator. This reduces the chance of a correct diagnosis in detecting acute lymphoblastic leukemia, a potentially fatal blood cancer if left untreated.

The study developed and presented an automated method for identifying and classifying leukocytes using microscopic images of peripheral blood smears. The proposed neural random threshold classifier achieved a recognition rate of 98.3% when the data was divided into 80% training set and 20% test set. The proposed system can be implemented as a computational tool to detect other diseases in which blood cells undergo changes, such as Covid-19. This will eliminate the subjective factor that is invariably inherent in the case of assessment by experts, focusing only on their experience. Therefore, the system is important for conducting expert assessments in medical diagnostics, as part of a decision support system that reduces the likelihood of risks of incorrect diagnosis of diseases.

Analysis of the capabilities of the Internet of Things in monitoring the physiological state and location of personnel on an offshore oil platform.
The possibilities of using the Internet of Things (IoT) to ensure personnel safety on an offshore oil platform were explored. For this purpose, IoT applications and technologies for monitoring the physiological state and location of personnel are analyzed. The use of cloud technologies, big data technologies and artificial intelligence for the development of systems that allow monitoring and, if necessary, making appropriate decisions through systematic monitoring of personnel status is considered. The basis for decision making is expert assessments of deviations of real-time parameter values from the norm. Practical problems associated with the use of Internet of Things technologies in various areas of healthcare are presented.

Methodological approaches to intelligent management of human factors on offshore oil and gas platforms.
The problems of increasing the efficiency of managing the safety and health of shift workers in the offshore oil and gas industry through the prism of the human factor have been studied. The environmental features, hazards and risks, working and professional conditions in the offshore sector are taken into account. The concept of a person-centered approach to personnel safety and health management is proposed, which involves the inclusion of employees in the management loop as the main component of their contextual environment. This involves constant remote monitoring of vital health indicators of employees and, at the same time, parameters of the context-dependent environment of each of them, as well as an expert assessment of the deviation of these parameters from the norm.

Decision support in a remote health monitoring system for shift workers on an offshore oil platform.
A methodological approach to the synthesis of solutions in a geographically distributed intelligent health management system for oil workers working in the maritime industry is proposed. A functional model of the health management system for workers employed on offshore oil platforms has been developed and implemented in three stages: monitoring and assessment of health indicators and environmental parameters of each employee, as well as decision-making. These interacting operations integrate layers of a distributed intelligent healthcare management system.

Appropriate approaches to implementing decision support processes are presented and one of the possible methods for assessing generated data and making decisions using fuzzy pattern recognition is described. Models of a fuzzy ideal image and a fuzzy real image of the employee’s health state have been developed and an algorithm for expert assessment of the deviation of the formed medical indicators from the norm is described.

Expert assessment of engineering and planning solutions to improve the safety of vulnerable road users in Ukraine.
The study is devoted to the analysis of the main methodological provisions for conducting an expert assessment (audit) of engineering and planning solutions, organizational and management measures to ensure road safety (RS) for vulnerable road users.

The results of the study are presented in the form of the authors’ opinions on the following problems (areas of activity) in the field of road safety, which are subject to expert assessment:
– the real level of road accidents among the main categories of vulnerable road users in Ukraine;
– modern scientific, methodological and engineering planning approaches to the formation of individual elements of an effective and safe transport infrastructure, as well as progressive traffic management systems (TS);
– the possibilities and feasibility of implementing engineering and planning solutions, organizational and management measures to improve road safety (RS).