User weighting

List of current features being used in user weighting and their associated parameter weighting. These weights were determined based on parameter optimization. The strategy for generating these features and corresponding weights are all open to change.

return: 0.9

return_s: 2.8

return_w: 0.7

accuracy: 1.7

accuracy_s: 0.6

all_features: 1.5

all_features_s: 0.9

all_features = accuracy * return * num_tweets

Last updated