The results section MTD is split into two: first a statistical assessment of the results is presented, together with a comparison of the other state-of-the art algorithms, followed by a human topic identification and classification test. They analyze 1000 reviews for 50 types of electronic products and 50 types of non-electronic products, with AMC finding a 5% better topic coherence than LTM, a tweaked version of LDA, which is 10% less coherent than AMC. In the human topic labeling test, judges decided whether a certain topic was coherent or not or whether a word really belonged to a certain topic. As this is a qualitative test, it is hard to quantify the results, but the AMC performed visibly better than LDA and slightly better or on par with LTM. In conclusion MTD showed that AMD has a vast array of potential applications, on top comment selection optimization, such as mining news for event prediction or mining scientific research articles to aid medicine / complex decisions.