Error analysis in big data analysis

Abstract

Practice errors in medicine are among the situations that disrupt patient-physician communication, reduce work efficiency, and eventually lead to the formation of an atmosphere of distrust. Various aspects of the subject have been examined in the symposiums we have held with Prof Dr. Ethem Geçim for 14 years. We shared our information at medikolegalduzlem.com. Big Data analysis is important in system audits. It is inevitable to reach erroneous results if the data belonging to these are made carelessly and negligently. The main problems encountered in big data analysis are: 1. Dependency: Computer dependency is at the forefront of algorithm and process monitoring. 2. Chaotic situation: The problem of systematic change is observed in the supervised learning machine. In the unsupervised learning machine, the definition of the identity of grouping is problematic. In the learning machine system of the reward-punishment system, there is a case study based on random data, not systematic. It is very difficult to reach an inductive conclusion from this situation. There is absolutely a need to organize a constant, standard data flow. 3. Loss of control: Steps that need to be checked are skipped in big data monitoring. Injustice occurs. 4. Waiver of rights: The expert is content with the data given by the machine as a dependent. There is a problem with the development of innovative processes. In that case, it is useful to pay attention to the following points in the regulation on data ethics: 1. Identity: In the big data analysis of the expert, it is important to classify and group the personal data and to accept it in court. In this respect, it is necessary to control whether the expert is legitimate, sufficient, and effective. 2. Confidentiality: The problem of privacy of the person being examined is important. The approach of the telemedicine application, which we call image-based analysis, to this problem is subject to legal and ethical evaluation. 3. Ownership: There are inadequacies in the regulation of the infrastructure applied in the field of health. The purchase of each tool and its accessible application also needs regulation. 4. Reputation: Audit problematic of non-standard practices of practitioners is evident. As a result, there is a need to evaluate error definition, error measurement, error effect, and error tolerability in big data analysis.