Relationship among similarity coefficients used in randomly amplified polymorphic DNA RAPD analysis/
Palma, Arthelo P.
Relationship among similarity coefficients used in randomly amplified polymorphic DNA RAPD analysis/ Arthelo P. Palma. - 2006 - 123 leaves.
Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2006
Contingency-based similarity coefficients like simple matching, Russel and Rao, Rogers and Tanimoto, Ochiai II, Sokal and Sneath, Jaccard, Anderberg, Dice, Ochiai and Kulczynski II were studied based on their properties, the similarity matrices they generated and the resulting dendograms. Some of these coefficients were widely used in divergency studies and in Randomly Amplified Polymorphic DNA analysis in particular. The similarity coefficients that do not consider negative co-occurrence like Jaccard, Anderberg and Dice, based on their properties, Spearman rank correlation and Fowlkes-Mallows index, were found to produce similar dendrograms. Similarly, the similarity coefficients that consider both negative and positive co-occurrence such as the Simple Matching, Rogers and Tanimoto, and Sokal and Sneath showed closely related and comparable output. It is important to know the comparability of the output of the similarity coefficients, for it implies no significant different of results. The Russell and Rao deviated from the group because it did not satisfy some desirable properties such as reflexivity and definiteness. Some other considered desirable properties such as reflexivity and definiteness. Some other considered desirable properties are standardization, reflexivity, symmetry, definiteness and monotonicity. Except for the Russel and Rao, the rest of the similarity coefficient satisfy these properties however, not all of them tend to give some result in performance. This implies that more mathematical properties need to be uncovered to find out additional information about the similarity coefficients used in clustering and in RAPD analysis. On the other hand, single linkage clustering algorithm was found to give better information than the average and complete linkage methods.
Undergraduate Thesis --AMAT200,
Relationship among similarity coefficients used in randomly amplified polymorphic DNA RAPD analysis/ Arthelo P. Palma. - 2006 - 123 leaves.
Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2006
Contingency-based similarity coefficients like simple matching, Russel and Rao, Rogers and Tanimoto, Ochiai II, Sokal and Sneath, Jaccard, Anderberg, Dice, Ochiai and Kulczynski II were studied based on their properties, the similarity matrices they generated and the resulting dendograms. Some of these coefficients were widely used in divergency studies and in Randomly Amplified Polymorphic DNA analysis in particular. The similarity coefficients that do not consider negative co-occurrence like Jaccard, Anderberg and Dice, based on their properties, Spearman rank correlation and Fowlkes-Mallows index, were found to produce similar dendrograms. Similarly, the similarity coefficients that consider both negative and positive co-occurrence such as the Simple Matching, Rogers and Tanimoto, and Sokal and Sneath showed closely related and comparable output. It is important to know the comparability of the output of the similarity coefficients, for it implies no significant different of results. The Russell and Rao deviated from the group because it did not satisfy some desirable properties such as reflexivity and definiteness. Some other considered desirable properties such as reflexivity and definiteness. Some other considered desirable properties are standardization, reflexivity, symmetry, definiteness and monotonicity. Except for the Russel and Rao, the rest of the similarity coefficient satisfy these properties however, not all of them tend to give some result in performance. This implies that more mathematical properties need to be uncovered to find out additional information about the similarity coefficients used in clustering and in RAPD analysis. On the other hand, single linkage clustering algorithm was found to give better information than the average and complete linkage methods.
Undergraduate Thesis --AMAT200,