def df_pca(df, n_comp=2): rd_df = normalize_dataframe(df) rd = PCA(n_components=n_comp) rd.fit(caracteristicas_df) caracteristicas_rd = rd.transform(rd_df) caracteristicas_rd_df = pd.DataFrame(caracteristicas_rd) return caracteristicas_rd_df
def df_mds(df, n_comp = 2, n_jobs = 1, max_iter = 1000): rd_df = normalize_dataframe(df) rd = MDS(n_components=n_comp, max_iter=max_iter, metric=True, n_jobs=n_jobs, random_state=2019) # rd.fit(caracteristicas_df[:1000]) caracteristicas_rd = rd.fit_transform(caracteristicas_df) caracteristicas_rd_df = pd.DataFrame(caracteristicas_rd) return caracteristicas_rd_df
def df_isomap(df, n_comp = 2, n_jobs = 1, n_neighbors = 5, max_iter = 1000): rd_df = normalize_dataframe(df) rd = Isomap(n_components=n_comp, n_neighbors = n_neighbors, max_iter = max_iter ) rd.fit(caracteristicas_df) caracteristicas_rd = rd.transform(rd_df) caracteristicas_rd_df = pd.DataFrame(caracteristicas_rd) return caracteristicas_rd_df
def df_kernel_pca(df, n_comp = 2, n_jobs = 1): rd_df = normalize_dataframe(df) rd = KernelPCA(kernel="rbf",n_components=n_comp, gamma=None, fit_inverse_transform=False, random_state = 2019, n_jobs=n_jobs) rd.fit(caracteristicas_df) caracteristicas_rd = rd.transform(rd_df) caracteristicas_rd_df = pd.DataFrame(caracteristicas_rd) return caracteristicas_rd_df
def df_svd(df, n_comp = 2, max_iter = 10): rd_df = normalize_dataframe(df) rd = TruncatedSVD(n_components=n_comp,algorithm='randomized', random_state=2019, n_iter=max_iter) rd.fit(caracteristicas_df) caracteristicas_rd = rd.transform(rd_df) caracteristicas_rd_df = pd.DataFrame(caracteristicas_rd) return caracteristicas_rd_df