特征组合

从元素数据组合为组合数据。

>>> from featurebox.featurizers.atom.mapper import AtomTableMap, AtomJsonMap
>>> data_map = AtomJsonMap(search_tp="name", n_jobs=1)
>>> wa = WeightedAverage(data_map, n_jobs=1,return_type="df")
>>> x3 = [{"H": 2, "Pd": 1},{"He":1,"Al":4}]
>>> wa.fit_transform(x3) # or
>>> wa.fit_transform(structure_list)

         0         1         2   ...        13        14        15
0  0.422068  0.360958  0.201433  ... -0.459164 -0.064783 -0.250939
1  0.007163 -0.471498 -0.072860  ...  0.206306 -0.041006  0.055843

[2 rows x 16 columns]


>>> wa.set_feature_labels(["fea_{}".format(_) for _ in range(16)])
>>> wa.fit_transform(x3)


      fea_0     fea_1     fea_2  ...    fea_13    fea_14    fea_15
0  0.422068  0.360958  0.201433  ... -0.459164 -0.064783 -0.250939
1  0.007163 -0.471498 -0.072860  ...  0.206306 -0.041006  0.055843

[2 rows x 16 columns]