classification.py 文件源码

def featuresByLSA(features,ncomponents=100):
    svd = TruncatedSVD(n_components=ncomponents)
    normalizer =  Normalizer(copy=False)
    lsa = make_pipeline(svd, normalizer)
    dtm_lsa = lsa.fit_transform(features)
    return dtm_lsa