We consider learning in a transductive setting using instance-based learning (k-NN) and present a method for constructing a data-dependent distance "metric" using both labeled training data as well as available unlabeled data (that is to be classified by the model). This new data-driven ...
Finally, we considered two modified formulations of widely used indices, tested on sessile oak in Turkey [30], integrating both diameter and total height dimensions on the same competition metric. Indices abbreviation, formulation and references can be found in Table 2. Table 2. Mathematical ...