The evaluations represent the concordance and the disagreement of class area distribution from the use of heuristic judgements and the objective datasets.
Choosing an objective testing dataset is the most difficult, especially when the datasets used in the original publications are not equivalent.
Figure 12 shows the global precipitation and evaporation (30-year average for 1981 2010) as functions of coastline length (Ogino et al. 2017) calculated from an objective analysis dataset called the Japanese 55-year Reanalysis (JRA−55) for 1981 2010 (Kobayashi et al. 2015).
This section describes in details the dataset, objective metrics, and performance analysis used in our benchmark.
To make a relatively objective and balanced dataset, for the BDB, protein targets and their associated compound affinities data were selected based on the following criteria: (1).
It addresses the following objective: given a dataset of n units described by m attributes and a positive integer k (the number of clusters), group the n units into k clusters so that the total sum of the distances of each unit to its nearest cluster center is minimized.
For objective 2, 12 datasets from the literature (11 from objective 1) containing SGR, FR, final BW, body lipid, protein, ash, moisture, and energy (kJ/g) were used.
For objective 1, 19 datasets from the literature containing initial BW, FR and specific growth rate (SGR; % BW increase per day) were used.
We are in the early phases of planning a curation strategy, but our general objectives for the dataset are to digitize it, add metadata, convert sampling data to actionable formats, and make it all public.
