Researchers expanded an efficiency based classifier system by using fuzzy rules to generate a rule set for predicting yarn strength. The fuzzy efficiency based classifier system enhances the original learning classifier algorithm of Goldberg by introducing several rule efficiencies into the learning strategy of the system. As a method to predict the spinnability and strength of yarn with a set of if-then rules, a test confirmed that a three fuzzy set implementation of yarn strength was a good compromise between accuracy and generalism, resulting in a total prediction accuracy of 92 percent. The generated rules enabled a translation toward comprehensible linguistic rules containing qualitative information about the fiber to yarn manufacturing process. Each rule contained a measure of its applicability that permitted the user to distinguish important, accurate rules from unimportant or faulty rules. The results from a comparison of a learn file and a separate test file revealed a danger of overfitting when applying higher k values. 12 refs.