The data sets used in this work can be divided into four groups by predictive task: prediction of protein stability change on mutation, prediction of protein protein and protein DNA affinity change on mutation.
In this section, we test our support for one sense-making task, prediction, by evaluating the predict missing links algorithm.
The fourth column of each task prediction contains the confidence score (Conf).
In contrast, in the OOO task, even our across-task predictions are as good as (even marginally better, p = 0.092, than) the within-task predictions of the GP classifier algorithm.
Furthermore, the subjective distributions we extracted from the familiarity task also provided across-task predictions in the OOO task that were as accurate as within-task predictions in that task (p = 0.84).
Importantly, the GP classifier directly fits subjects' stimulus-to-response mappings without extracting underlying subjective distributions and thus has no way to provide across-task predictions.
Remarkably, within-task predictions for the familiarity task are very close to an expected upper bound that can be computed based on subjects' consistency [ 25].
We compared two methods for predicting muscle activities and postures, based on only the distribution of keyboard/mouse/idle tasks (task-based predictions) or on a comprehensive set of 104 task, questionnaire, workstation, and anthropometric parameters (expanded model predictions).
In the initial single-task prediction context, a task is defined as the learning of a personality trait.
The target values y of the tasks were calculated using the standard multi-task prediction function (6), which means that the target values do not contain label noise.
In addition, in contrast to most of the existing methods, which applied instance selection to classification tasks (discrete prediction), the proposed approach is used to obtain instance selection methods for regression tasks (prediction of continuous values).
