前向特征选择

with open('30second_genes_list.txt', 'w') as f:

#    declare the method

    selector = FwdSelect()

#    fit the selector with the cell data

    selector.fit(df.values)

#    rank the features

    idx = selector.ordered_features

#    get top 20 features

    ls = [gene_list[e] for e in idx][:20]

#    write down selected genes into a file

    f.write('\n'.join(ls))