rule1 :title_exact == 1.0 rule2 :author_trigram >= 0.7 AND jaccard_title >= 0.75 rule2.1 :author_trigram >= 0.75 AND date == 1.0 rule2.3 :author_trigram >= 0.75 AND date == 1.0 AND pages == 1.0 rule3 :date == 1.0 AND author_trigram >= 0.7 AND trigram_title >= 0.5 rule4 :jaccard_title >= 0.5 AND date == 1.0 AND pages == 1.0 rule5 :trigram_title >= 0.8 AND date == 1.0 rule9 :trigram_title >= 0.8 AND pages == 1.0 MATCHER1 : rule1 RESULT : P = 1.0, R=.58 MATCHER2 : rule2.1 RESULT : P = 0.87, R = 0.61 MATCHER : rule2.1 OR rule1 RESULT : P = 0.97, R = 0.86 MATCHER : rule1 OR rule2.3 OR rule2 RESULT : P = 0.97, R = 0.87 MATCHER : rule1 or rule2.3 or rule5 RESULT : P = 0.97, R = 0.88 MATCHER : rule1 or rule2.3 or rule3 RESULT : P = 0.96, R = 0.90 MATCHER : rule1 or rule2.3 or rule4 RESULT : P = 0.95, R = 0.89 MATCHER : rule1 or rule2.3 or rule9 RESULT : P = 0.97, R = 0.89 MATCHER : rule1 or rule2.3 or rule5 or rule2 RESULT : P = 0.97, R = 0.89 MATCHER : rule1 or rule2.3 or rule5 or rule3 RESULT : P = 0.96, R = 0.92 MATCHER : rule1 or rule2.3 or rule5 or rule3 or rule4 RESULT : P = 0.94, R = 0.93 MATCHER : rule1 or rule2.3 or rule2 or rule3 or rule5 RESULT : P = 0.96, R = 0.93 OBSERVATIONS: Below are some of our observations we made comparing the results and manually scanning the data. The precision and recall could not be improved further because: 1) book title differs by one word eg. Fodor's In Focus Charleston: with Hilton Head & the Lowcountry vs Fodor's In Focus Savannah: with Hilton Head & the Lowcountry (unfortunately sometimes different books by the same author are published on the same date). 2) book title differs by the words for all books that belong to some series. eg. Creed (Unfinished Heroes) (Volume 2) in Amazon vs Creed in BarnesandNoble