edu/per/does/Addendum on normalizedgain. pdf. Sriyati, S. , Rustaman, N. , and Zainul, A. 2010. Then, each new data element must be converted driving each transformation to examine its distance to matlab mean of each category for instance for matlab two category case, matlab new data point could need to be converted by matlab category 1 transformation and then compared to mu 0 and matlab new data aspect could even have to be converted by matlab category 2 transformation and then compared to mu 1 . The difference between Case 1 and Case 2 i. e. the change between using matlab Euclidean and Mahalanobis distance may also be seen in matlab example beneath. As will also be seen from matlab instance above, matlab Mahalanobis distance takes under consideration matlab distribution engineering matlab data points, while matlab Euclidean distance may treat matlab data as however matlab has engineering spherical distribution. Thus, matlab Mahalanobis distance applies for matlab more standard classification in Case 2, whereas matlab Euclidean distance applies to matlab particular case in Case 1 in which matlab data distribution is thought to be round.