Diagnostics

Diagnosis (Greek: d?????s?, from dia-d?a "out-split", and ???s? gnosi to learn, knowledge ") is to identify the nature of anything, either due to exclusion or methods of analysis. Diagnosis is used in different fields, with a slightly different reality on the application of logic and experience to determine the cause and effect relationship. Below are given as examples and tools used by the respective specialized in medicine, science, technology business. Diagnosis is also used in other industries and professions to determine the cause of symptoms, mitigations for problems or solutions to problems.

Is a subfield of artificial intelligence, diagnosis is associated with the development of algorithms and techniques that can determine the behavior of a system is correct. If the system does not function properly, the algorithm should be able to determine, as accurately as possible, but part of the system is not, and the type of error it is facing. Calculation is based on observations, which provide information about the current behavior.

The diagnosis of participants also mentioned the answer to the question whether the system has problems or not, and the process of calculating the answer. The word comes from the context of health care, where diagnosis is the process of determining a disease by its symptoms.

Specialist Diagnosis Expert diagnosis (or diagnostic expert system) is based on experience with the system. Using experience, mapping built efficiently associates observations to diagnose respectively.

Experience can be provided:
* According to a human operator. In this case, knowledge of man must be translated into a language the computer.

* Follow the example of the behavior system. In this case, for example, be classified as right or wrong (and, in the case later, by type of error). Machine learning methods are then used to generalize from examples. The main disadvantage of this method is

* Difficulties have been professional. Experts available only after a long time use of the system (or similar system). Therefore, these methods are not suitable for the safety or mission critical systems (such as a nuclear power, or operating a robot in space). Furthermore, the expertise to buy never can ensure are completed. In the case of a previously unseen behavior occurs, leading to an unexpected observation, unable to give a diagnosis.

* The complexity of the study. The off-line process of building an expert system can require a large amount of time and computer memory.

* The size of the final expert system. Expert system to map any observations to diagnose, it will in some cases requires a large amount of storage space.

* The lack of strong. If even a small change is made on the system, the process of building expert systems must be repeated.

A slightly different approach is to build a system of experts from a model of the system rather than directly from a professional. For example, calculation of a diagnoser for diagnosis of discrete event systems. This approach can be viewed as model-based, but it benefited from a number of advantages and disadvantages of being a number of approaches systems specialist.