For example, 16.72% of molecules of DS1 are clustered as singletons when Wards method is applied on ALOGP fingerprint with number of clusters equal to 1000 clusters.
In the tests on data from 20 clusters we started from a lower number of initial clusters (equal to 10), in order to check the ability of our algorithm to detect the correct number of clusters.
This is achieved without any filtering, using a search template of size 21 × 21, and a number of clusters equal to 1/200 of the total number of training patterns.
As our hypothesis bases on the effect of the soil characteristics on the corrosion process of the coins, we expected to obtain "a series" of clusters, equal to the number of coins, each one connected to bigger clusters as a function of the different soil characteristics.
Finally, the most accurate E_type of Spectral cases is shown in Fig. 6e; using PCA as filtering method, a search template of size 17 × 17, and a number of clusters equal to 1/200 of the total number of training patterns has led to the most accurate result obtained through application of Spectral clustering.
The least accurate E_type of FILTERSIM using Spectral clustering method is demonstrated in Fig. 6d where it comes from KNN as filtering method, a search template of size 29 × 29, and a number of clusters equal to 1/200 of the total number of training patterns.
As it is clear in Table 3, the least accurate result of K-means cases is obtained using KNN as filtering method, a search template of size 17 × 17, and a number of clusters equal to 1/800 of the total number of training patterns (i.e., a total number of 7,176 scanned patterns).
The strategy adopted by the PAM-SLIM algorithm, for example, is based on the assumption that each level of the Slim-tree indirectly divides the data space into a number of clusters equal to the number of elements stored at each level.
The k-means algorithm was parameterized with a number of final clusters equal to 2, 2 random observations to choose the initial cluster centroid positions, 30 replicates and with the L1 distance to calculate the distance between centroid clusters.
The stringency was increased to identify significant SNPs that clustered at least two accessions with low or high SGA levels, with mean difference between SNP clusters equal to or greater than 5 logarithm units (Table 8).
