A selection of 12 correlations between 2D fingerprints using the average score of evaluation method BEDROC 20) for the three collection of data sets: MUV (open triangles), DUD (open circles), and ChEMBL (filled circles).
The item loading for the one-factor solution for the PCA method and average score for each item is shown in Table 2.
In our simulation tests of three genomes with 85 genes each, and branch lengths of 50, 20 and 20 (roughly corresponding to the branches leading to C. reinhardtii, C. vulgaris and N. olivacea; see Methods), the average score for ancestral gene orders computed by GRAPPA was only about 7% less than the true scores.
A selection of 12 correlations between evaluation methods using the average score of 2D fingerprint ECFP4 for the three collection of data sets: MUV (open triangles), DUD (open circles), and ChEMBL (filled circles).
In this analysis, there were 3057 FAexpTRL from Pm1, 2, 3, 4, 5, 7, 8, 9 and 10 which were in the top 10% (red zone) for the average monocot % alignment category and which also had significant alignments with all 4 monocot test databases (i.e. an average score of 1, see Materials and Methods).
The average score (computed as the sum of scores for each method divided by the total number of methods m) was scaled from 0-10 by multiplying by 10; the higher the score, the better the agreement among the four methods.
Therefore, the average score indicates the performance of the recommendation methods.
The average score (8.1out of 10) of the students was independent of the method used, but more students did the practice with non-attendant method (57 vs 26%).
To convert this into an election method, count each candidate's stars or points, and the winner is the one with the highest average score (or the highest total score, if voters are allowed to leave some candidates unrated).
