Page 106 - Demo

%u062c%u0645%u064a%u0639 %u0627%u0644%u062d%u0642%u0648%u0642 %u0645%u062d%u0641%u0648%u0638%u0629 %u0640 %u0627%u0625%u0644%u0639%u062a%u062f%u0627%u0621 %u0639%u0649%u0644 %u062d%u0642 %u0627%u0645%u0644%u0624%u0644%u0641 %u0628%u0627%u0644%u0646%u0633%u062e %u0623%u0648 %u0627%u0644%u0637%u0628%u0627%u0639%u0629 %u064a%u0639%u0631%u0636 %u0641%u0627%u0639%u0644%u0647 %u0644%u0644%u0645%u0633%u0627%u0626%u0644%u0629 %u0627%u0644%u0642%u0627%u0646%u0648%u0646%u064a%u0629106To generate a rule while(true) find the best predicate pif foil-gain(p) > threshold then add p to current rule else break How to Learn-One-Rule? Start with the most general rule possible: condition = empty, Adding new attributes by adopting a greedy depth-first strategy, Picks the one that most improves the rule quality Rule-Quality measures: consider both coverage and accuracy. Foil-gain (in FOIL & RIPPER): assesses info_gain by extending condition, favors rules that have high accuracy and cover many positive tuples pos' posFOIL _ Gain = pos' (log 2 %u2212 log 2 )pos'+neg' pos+negRule pruning based on an independent set of test tuples FOIL_Prune(R)= pos%u2212negpos+neg