Expert diagnosis of computer systems using neuro-fuzzy knowledge base

The functional diagnosis expert system based on knowledge base in the form of neuro-fuzzy network was proposed in this paper. Current values of diagnostic parameters are measured by sensors. The hybrid expert diagnostic system with neuro-fuzzy network knowledge base supports decisions in the situation when the diagnosis algorithm is not known and is formed from the initial data in the form of production rules. Sensors use to automate the process of accumulation of knowledge in the expert system. The need to take decisions in real time results the number of trees corresponding to incoming data, equal to the number of samples during the observation period. The problem of large amounts of data in determining the technical condition of the complex technical object is solved by using this data as a training sample for the knowledge base. The expert diagnosis system for assessing the PC performance was considered how the example.