Top n best

This is a sample checking the best n results in target generation.

Note:

Only valid in store=True.

if __name__ == "__main__":
    from sklearn.datasets import load_iris
    from bgp.skflow import SymbolLearning
    from sklearn import metrics
    from sklearn.utils import shuffle

    data = load_iris()
    x = data["data"]
    y = data["target"]
    c = [1, 2, 3]

    sl = SymbolLearning(loop="MultiMutateLoop", pop=50,
                        re_hall=3,
                        gen=3, random_state=1,
                        classification=True,
                        scoring=[metrics.accuracy_score, ], score_pen=[1, ],
                        store=True,
                        )

    sl.fit(x, y, c=c)
    score = sl.score(x, y, "r2")
    top_n = sl.loop.top_n(10)
    print(sl.expr)

where the top n is a table (Pandas.DataFrame object), as following: