ABSTRACT
Resolving anomalies is essential for many tasks like diagnostic reasoning, scientific discovery and everyday problem solving. Abductive reasoning is viewed as an inference to the best explanation for a given set of observations or data. The process of abduction is viewed as the sequential comprehension and integration of evidence into a single situation model that represents the current best explanation. If the generated explanation or part of it is inconsistent with any other hypothesis or data in the situation model an anomaly has occurred. As a consequence, the model must be updated by either finding an alternative explanation or by changing an explanation for the old data. If anomalous data occur the model responds by rejecting all but one of the anomalous explanations and then constructing alternative explanations for the data that were left unexplained by the rejection(s).
