Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Table 1.

Respondents in each mode and each Round by country.

More »

Table 1 Expand

Fig 1.

Depiction of the machine learning procedure.

The workflow illustrates the sequential steps used in the analysis, including data preprocessing, feature extraction, model training, validation, and evaluation. Arrows indicate the steps of the procedure and the iterative process of model refinement.

More »

Fig 1 Expand

Fig 2.

Metrics for per-country models among the countries with different administration modes between Rounds 10 and Rounds 8-9.

The figure presents model performance across countries that experienced a change in survey administration mode between rounds, using metrics of test accuracy and F1-score relative to random accuracy. See Table 1 for a key to country abbreviations.

More »

Fig 2 Expand

Fig 3.

Metrics for per-country models among the countries with identical administration modes between Rounds 10 and Rounds 8-9.

The figure presents model performance across countries that experienced no change in survey administration mode between rounds, using metrics of test accuracy and F1-score relative to random accuracy. See Table 1 for a key to country abbreviations.

More »

Fig 3 Expand

Fig 4.

Metrics for models with all data from a given country in the unseen data, for countries with differing administration modes.

The figure presents model performance across countries that experienced a change in survey administration mode between rounds, using metrics of test accuracy and F1-score relative to random accuracy, when all of a given country’s data was in the unseen data and the model was trained on the remaining countries. See Table 1 for a key to country abbreviations.

More »

Fig 4 Expand

Fig 5.

Metrics for models with all data from a given country in the unseen data, for countries with identical administration modes.

The figure presents model performance across countries that experienced no change in survey administration mode between rounds, using metrics of test accuracy and F1-score relative to random accuracy, when all of a given country’s data was in the unseen data and the model was trained on the remaining countries. See Table 1 for a key to country abbreviations.

More »

Fig 5 Expand