In order to test a fingerprint's ability to order molecules by structural similarity, we created two distinct benchmark datasets of thousands of series of 5 molecules, each series containing a reference molecule and an ordered arrangement of 4 molecules.
They may be inadequate, however, for use in population genomics/genetics, involving large datasets of thousands of individuals.
At the same time, one could argue that the 70-gene set, for example, may not necessarily represent the 'best' or optimal set of predictors, and many have advocated using training datasets of thousands, rather than tens, of patient profiles for developing more robust prognostic signatures.
Data for this study was gathered over the course of 3 years at several art institutions aiming to gather a representative dataset of 20th century paintings.
Assuming that a maximum of 20%% of the CPM values at the high end of the distribution were due to outliers, these values were discarded by building two datasets each: the 80th, 85th, 90th, 95th and 99th quantiles (Q80 Q99), and the 80, 85, 90, 95, and 99%% trimmed sums (TS80 TS99).
It is also interesting to note that in some datasets of TS/1 (i.e., 2nd, 3rd and 4th) and of TS/2 (i.e., 4th, 5th and 6th) the optimal λ precedes a rapid increase in the number of communities.
According to the dataset of the 6th census in 2010, the average natural growth rate of the population in Xinjiang, Tibet, Ningxia, Gansu and Qinghai in the northwest registered 7.86%, 1.63 times of the national average of 4.81% during the same period.
A comparison of pairwise divergences within the four-species datasets based on 1st and 2nd versus 3rd positions suggested 14 times higher substitution rate at 3rd positions equivalent to 3.44% per My.
Fly L3 larvae: two datasets of wild-type (Oregon R) early 3rd instar larvae (E. L. Garcia and A. G. Matera, unpublished data).
For each of the six permuted subsets of Datasets 4, the 90th percentile penalized t-statistic, SAM penalized t-statistic, and Z(Iavg) had similar R2 values when correlated with the "gold standard" t-statistic, although the SAM statistic did perform poorly for the noisiest subset of Dataset 4 (#3 in Table 2) with an R2 value of only 0.70.
1-NN, C4.5 and NB are always bottom of the ranking for each of the seven datasets (7th, 6th and 5th position) and they lose a lot of accuracy when the number of classes increases (from 57 to 389), especially 1-NN and C4.5 which, respectively, decrease from (59.9) to 28.75 % and from 75.0 to 45.8 %, while NB decreases only from 85.2 to 74.6%%.
