Majority voting

Model

The Majority voting model returns the most voted label among the $K$ labels, for each task:

\[\DeclareMathOperator*{\argmax}{argmax} \hat y_i = \argmax_{k\in[K]} \sum_{j\in [n_\texttt{worker}]} \unicode{x1D7D9}_{\{y_i^{(j)}=k\}}.\]

With peerannot in a terminal located in the directory of answers, the MV model can be used as follows.

peerannnot aggregate . --strategy MV --answers-file answers.json

Note that by default, if the answers are in a file names answers.json the --answers-file argument can be omitted.

Import the aggregation model in the current session

from peerannot.models import MV

Assuming the answers are in a dictionary names answers then run:

mv = MV(answers, n_classes)
yhat = mv.get_answers()

Note that the majority voting aggregation produces hard labels (Dirac distributions).

MV model class inherits from CrowdModel

__init__(answers, n_classes,**kwargs)

Parameters:

answers:(dict) Dictionary of workers answers with format

          {
              task0: {worker0: label, worker1: label},
              task1: {worker1: label}
          }

n_classes: (int) Number of possible classes
kwargs: (dict) Dictionary that can contain path_remove to remove tasks identified from the WAUM or another method.

compute_baseline()

For each given task, computes the number of votes for each label.

get_answers()

Returns the most voted class from the baseline. In case of equalities, a random choice is applied.