Extended API

Semi-Supervised Learning

Use when you have many rows but only some have labels. Mark unlabeled rows with -1 in the target column.

SemiSupervisedClassifier

python
from fling import SemiSupervisedClassifier

df["label"] = df["label"].fillna(-1).astype(int)  # -1 = unlabeled

ssc = SemiSupervisedClassifier(df, target="label", kernel="rbf", alpha=0.2)
ssc.train()
all_labels = ssc.predict()

LabelPropagator

python
from fling import LabelPropagator

lp = LabelPropagator(df, target="label")
lp.train()