FTRLProximal

FTRL-Proximal optimizer.

Parameters

• alpha

  Default → 0.05

• beta

  Default → 1.0

• l1

  Default → 0.0

• l2

  Default → 1.0

Attributes

• z (collections.defaultdict)

• n (collections.defaultdict)

Examples

from river import datasets
from river import evaluate
from river import linear_model
from river import metrics
from river import optim
from river import preprocessing

dataset = datasets.Phishing()
optimizer = optim.FTRLProximal()
model = (
    preprocessing.StandardScaler() |
    linear_model.LogisticRegression(optimizer)
)
metric = metrics.F1()

evaluate.progressive_val_score(dataset, model, metric)

F1: 87.56%

Methods

look_ahead

Updates a weight vector before a prediction is made.

Parameters: w (dict): A dictionary of weight parameters. The weights are modified in-place. Returns: The updated weights.

Parameters

• w — 'dict'

step

Updates a weight vector given a gradient.

Parameters

• w — 'dict | VectorLike'
• g — 'dict | VectorLike'

Returns

dict | VectorLike: The updated weights.

---

1. McMahan, H.B., Holt, G., Sculley, D., Young, M., Ebner, D., Grady, J., Nie, L., Phillips, T., Davydov, E., Golovin, D. and Chikkerur, S., 2013, August. Ad click prediction: a view from the trenches. In Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 1222-1230) ↩

2. Tensorflow's FtrlOptimizer ↩