An expert system for transient data analysis using a model based architecture developed with poplog

Signals, in the form of measured responses of an experiment, may be a confusing source of information. Whilst being a detailed and accurate description of a process, their presentation, in ´raw´ form, can initially be incomprehensible. However, the signal analyst is able to pick out patterns in the data and thus (with further processing) interpret the signal and make deductions on unknowns in the experiment. It is thought that an awareness of the experimental details enables the analyst to form a simple model of the process, which is then used to predict the general form of the responses. By working from the data, the analyst can improve the model in the sense that it becomes more specific to the experiment. The state of the model reflects the understanding of the process. When the analyst is confident of interpretation and deduction techniques, the model is sufficently complete to explain the experiment.