MonteCarloClassifierChain

Monte Carlo Sampling Classifier Chains.

Probabilistic Classifier Chains using Monte Carlo sampling, as described in ¹.

m samples are taken from the posterior distribution. Therefore we need a probabilistic interpretation of the output, and thus, this is a particular variety of ProbabilisticClassifierChain.

Parameters

• model

  Type → base.Classifier

• m

  Type → int

  Default → 10

  Number of samples to take from the posterior distribution.

• seed

  Type → int | None

  Default → None

  Random number generator seed for reproducibility.

Examples

from river import feature_selection
from river import linear_model
from river import metrics
from river import multioutput
from river import preprocessing
from river.datasets import synth

dataset = synth.Logical(seed=42, n_tiles=100)

model = multioutput.MonteCarloClassifierChain(
    model=linear_model.LogisticRegression(),
    m=10,
    seed=42
)

metric = metrics.multioutput.MicroAverage(metrics.Jaccard())

for x, y in dataset:
   y_pred = model.predict_one(x)
   y_pred = {k: y_pred.get(k, 0) for k in y}
   metric = metric.update(y, y_pred)
   model = model.learn_one(x, y)

metric

MicroAverage(Jaccard): 51.79%

Methods

learn_one

Update the model with a set of features x and the labels y.

Parameters

• x
• y
• kwargs

Returns

self

predict_one

Predict the labels of a set of features x.

Parameters

• x — 'dict'
• kwargs

Returns

dict[FeatureName, bool]: The predicted labels.

predict_proba_one

Predict the probability of each label appearing given dictionary of features x.

Parameters

• x
• kwargs

Returns

A dictionary that associates a probability which each label.

---

1. Read, J., Martino, L., & Luengo, D. (2014). Efficient monte carlo methods for multi-dimensional learning with classifier chains. Pattern Recognition, 47(3), 1535-1546. ↩