Search

Objective: The aim of this study is to introduce 23 different binary dissimilarity/distance coefficients determined for different values by using derived data sets and to evaluate them by revealing their properties. Material and Methods: In this study, the dissimilarity/distance coefficients put forward for binary data are considered. In the study, data were derived for 35 different values in the range of using the Phyton-random library. In the derivation of the data, firstly, which cell shown with will be assigned value, then the value to be assigned to the relevant cell was determined. 286 for, 815 for and 1000 different data sets for were used in the study. Results: Although the value range of all dissimilarity/distance coefficients for binary data is expected to be 0 to 1, this range is not valid for all coefficients. Out of 23 different coefficients, 12 take values in this range. According to the hierarchical clustering analysis, most of the dissimilarity/distance coefficients are similar. Conclusion: In general, the values of almost all coefficients increase from a fixed minimum to a fixed maximum as the samples become more dissimilar. However, the Sokal-Michener, Hamming and variance coefficients increase linearly with dissimilarity/distance smoothly for all values. The coefficient suggested by Sokal-Michener stands out among all the coefficients because the value range is 0-1 and is parallel to the dissimilarity/distance increase. Cosine, Hamming, Euclid I and Euclid II coefficients are affected by the number of and other coefficients are not. Therefore, it has been determined that a significant part of the dissimilarity/distance coefficients are independent of the sample size. [ABSTRACT FROM AUTHOR]