print('QOT Distribution - Object variables description') print(data.describe(include='object')) data.describe(include='object').to_csv( graphsDir + 'QOT Distribution - Object variables description.csv') print() print('Object variables values description') symbolic_vars = data.select_dtypes(include='object').columns for v in symbolic_vars: print(v, data[v].unique()) print() print('QOT Distribution (Object) - Histogram') rows, cols = (2, 3) fig, axs = plt.subplots(rows, cols, figsize=(cols * ds.HEIGHT, rows * ds.HEIGHT), squeeze=False) i, j = 0, 0 for n in range(6): counts = data[symbolic_vars[n]].value_counts() ds.bar_chart(counts.index.to_list(), counts.values, ax=axs[i, j], title='Histogram for %s' % symbolic_vars[n], xlabel=symbolic_vars[n], ylabel='nr records') i, j = (i + 1, 0) if (n + 1) % cols == 0 else (i, j + 1) plt.suptitle('QOT Distribution (Object) - Histogram') plt.savefig(graphsDir + 'QOT Distribution (Object) - Histogram')
f'DBSCAN metric={METRICS[n]} eps={distances[n]:.2f} k={k}', ax=axs[i,j]) else: print(k) mse.append(0) mae.append(0) sc.append(0) db.append(0) i, j = (i + 1, 0) if (n+1) % cols == 0 else (i, j + 1) plt.suptitle('QOT Clustering - Metric (Density-based) after PCA') plt.savefig(subDir + 'QOT Clustering - Metric (Density-based) after PCA') print('QOT Clustering - Metric (Density-based) MSE vs MAE vs SC vs DB after PCA') fig, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.bar_chart(METRICS, mse, title='DBSCAN MSE', xlabel='metric', ylabel='MSE', ax=ax[0, 0]) ds.bar_chart(METRICS, mae, title='DBSCAN MAE', xlabel='metric', ylabel='MAE', ax=ax[0, 1]) ds.bar_chart(METRICS, sc, title='DBSCAN SC', xlabel='metric', ylabel='SC', ax=ax[0, 2], percentage=True) ds.bar_chart(METRICS, db, title='DBSCAN DB', xlabel='metric', ylabel='DB', ax=ax[0, 3]) plt.suptitle('QOT Clustering - Metric (Density-based) MSE vs MAE vs SC vs DB after PCA') plt.savefig(subDir + 'QOT Clustering - Metric (Density-based) MSE vs MAE vs SC vs DB after PCA') print('QOT Clustering - Hierarchical after PCA') mse: list = [] mae: list = [] sc: list = [] db: list = [] rows, cols = ds.choose_grid(len(N_CLUSTERS)) _, axs = plt.subplots(rows, cols, figsize=(cols*5, rows*5), squeeze=False)
accuracy = 0 for clf in estimators: xvalues.append(clf) for trn_X, trn_y, tst_X, tst_y in zip(trn_x_lst, trn_y_lst, tst_x_lst, tst_y_lst): estimators[clf].fit(trn_X, trn_y) prd_tst = estimators[clf].predict(tst_X) accuracy += metrics.accuracy_score(tst_y, prd_tst) yvalues.append(accuracy / n_splits) plt.figure(figsize=(7, 7)) ds.bar_chart(xvalues, yvalues, title='Comparison of Naive Bayes Models', ylabel='accuracy', percentage=True) plt.suptitle(subDir + 'HFCR Naive Bayes - ' + key + ' - Comparison of Naive Bayes Models') plt.savefig(subDir + 'HFCR Naive Bayes - ' + key + ' - Comparison of Naive Bayes Models') if (offset == 1): break if (last_accuracy > best_accuracy and best_accuracy != -1): best_accuracy = last_accuracy last_accuracy = -1 count += offset offset -= 1 elif (best_accuracy == -1):
import os from pandas.plotting import register_matplotlib_converters graphsDir = './Results/Dimensionality/' if not os.path.exists(graphsDir): os.makedirs(graphsDir) register_matplotlib_converters() data = pd.read_csv('../Dataset/heart_failure_clinical_records_dataset.csv') print(data.shape) plt.figure(figsize=(4, 2)) values = {'nr records': data.shape[0], 'nr variables': data.shape[1]} ds.bar_chart(values.keys(), values.values(), edgecolor="teal", color="turquoise") plt.title('Nr of records vs nr of variables', color="teal") plt.savefig(graphsDir + 'HFCR Dimensionality - NrRecords.png') print(data.dtypes) data.dtypes.to_csv(graphsDir + 'HFCR Dimensionality - Types of variables.csv') cat_vars = data.select_dtypes(include='object') data[cat_vars.columns] = data.select_dtypes( ['object']).apply(lambda x: x.astype('category')) print(data.dtypes) data.dtypes.to_csv( graphsDir + 'HFCR Dimensionality - Types of variables - Improvement.csv') plt.figure(figsize=(20, 3))
def pca_function(data, subDir): data.pop('DEATH_EVENT') variables = data.columns.values eixo_x = 0 eixo_y = 4 eixo_z = 7 plt.figure() plt.xlabel(variables[eixo_y]) plt.ylabel(variables[eixo_z]) plt.scatter(data.iloc[:, eixo_y], data.iloc[:, eixo_z]) print('HFCR Feature Extraction - PCA') mean = (data.mean(axis=0)).tolist() centered_data = data - mean cov_mtx = centered_data.cov() eigvals, eigvecs = np.linalg.eig(cov_mtx) pca = PCA() pca.fit(centered_data) PC = pca.components_ var = pca.explained_variance_ # PLOT EXPLAINED VARIANCE RATIO fig = plt.figure(figsize=(4, 4)) plt.title('Explained variance ratio') plt.xlabel('PC') plt.ylabel('ratio') x_values = [str(i) for i in range(1, len(pca.components_) + 1)] bwidth = 0.5 ax = plt.gca() ax.set_xticklabels(x_values) ax.set_ylim(0.0, 1.0) ax.bar(x_values, pca.explained_variance_ratio_, width=bwidth) ax.plot(pca.explained_variance_ratio_) for i, v in enumerate(pca.explained_variance_ratio_): ax.text(i, v + 0.05, f'{v*100:.1f}', ha='center', fontweight='bold') plt.suptitle('HFCR Feature Extraction - PCA') plt.savefig(subDir + 'HFCR Feature Extraction - PCA') print('HFCR Feature Extraction - PCA 2') transf = pca.transform(data) _, axs = plt.subplots(1, 2, figsize=(2 * 5, 1 * 5), squeeze=False) axs[0, 0].set_xlabel(variables[eixo_y]) axs[0, 0].set_ylabel(variables[eixo_z]) axs[0, 0].scatter(data.iloc[:, eixo_y], data.iloc[:, eixo_z]) axs[0, 1].set_xlabel('PC1') axs[0, 1].set_ylabel('PC2') axs[0, 1].scatter(transf[:, 0], transf[:, 1]) plt.suptitle('HFCR Feature Extraction - PCA') plt.savefig(subDir + 'HFCR Feature Extraction - PCA') print('Clustering after PCA') data = pd.DataFrame(transf[:, :2], columns=['PC1', 'PC2']) eixo_x = 0 eixo_y = 1 N_CLUSTERS = [2, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29] rows, cols = ds.choose_grid(len(N_CLUSTERS)) print('HFCR Clustering - K-Means after PCA') mse: list = [] mae: list = [] sc: list = [] db: list = [] _, axs = plt.subplots(rows, cols, figsize=(cols * 5, rows * 5), squeeze=False) i, j = 0, 0 for n in range(len(N_CLUSTERS)): k = N_CLUSTERS[n] estimator = KMeans(n_clusters=k) estimator.fit(data) mse.append(estimator.inertia_) mae.append( ds.compute_mae(data.values, estimator.labels_, estimator.cluster_centers_)) sc.append(silhouette_score(data, estimator.labels_)) db.append(davies_bouldin_score(data, estimator.labels_)) ds.plot_clusters(data, eixo_x, eixo_y, estimator.labels_.astype(float), estimator.cluster_centers_, k, f'KMeans k={k}', ax=axs[i, j]) i, j = (i + 1, 0) if (n + 1) % cols == 0 else (i, j + 1) plt.suptitle('HFCR Clustering - K-Means after PCA') plt.savefig(subDir + 'HFCR Clustering - K-Means after PCA') print( 'HFCR Clustering - K-Means after PCA MSE vs MAE vs SC vs DB after PCA') fig, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.plot_line(N_CLUSTERS, mse, title='KMeans MSE', xlabel='k', ylabel='MSE', ax=ax[0, 0]) ds.plot_line(N_CLUSTERS, mae, title='KMeans MAE', xlabel='k', ylabel='MAE', ax=ax[0, 1]) ds.plot_line(N_CLUSTERS, sc, title='KMeans SC', xlabel='k', ylabel='SC', ax=ax[0, 2], percentage=True) ds.plot_line(N_CLUSTERS, db, title='KMeans DB', xlabel='k', ylabel='DB', ax=ax[0, 3]) plt.suptitle( 'HFCR Clustering - K-Means after PCA MSE vs MAE vs SC vs DB after PCA') plt.savefig( subDir + 'HFCR Clustering - K-Means after PCA MSE vs MAE vs SC vs DB after PCA') print('HFCR Clustering - Expectation-Maximization after PCA') mse: list = [] mae: list = [] sc: list = [] db: list = [] _, axs = plt.subplots(rows, cols, figsize=(cols * 5, rows * 5), squeeze=False) i, j = 0, 0 for n in range(len(N_CLUSTERS)): k = N_CLUSTERS[n] estimator = GaussianMixture(n_components=k) estimator.fit(data) labels = estimator.predict(data) mse.append(ds.compute_mse(data.values, labels, estimator.means_)) mae.append(ds.compute_mae(data.values, labels, estimator.means_)) sc.append(silhouette_score(data, labels)) db.append(davies_bouldin_score(data, labels)) ds.plot_clusters(data, eixo_x, eixo_y, labels.astype(float), estimator.means_, k, f'EM k={k}', ax=axs[i, j]) i, j = (i + 1, 0) if (n + 1) % cols == 0 else (i, j + 1) plt.suptitle('HFCR Clustering - Expectation-Maximization after PCA') plt.savefig(subDir + 'HFCR Clustering - Expectation-Maximization after PCA') print( 'HFCR Clustering - Expectation-Maximization MSE vs MAE vs SC vs DB after PCA' ) fig, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.plot_line(N_CLUSTERS, mse, title='EM MSE', xlabel='k', ylabel='MSE', ax=ax[0, 0]) ds.plot_line(N_CLUSTERS, mae, title='EM MAE', xlabel='k', ylabel='MAE', ax=ax[0, 1]) ds.plot_line(N_CLUSTERS, sc, title='EM SC', xlabel='k', ylabel='SC', ax=ax[0, 2], percentage=True) ds.plot_line(N_CLUSTERS, db, title='EM DB', xlabel='k', ylabel='DB', ax=ax[0, 3]) plt.suptitle( 'HFCR Clustering - Expectation-Maximization MSE vs MAE vs SC vs DB after PCA' ) plt.savefig( subDir + 'HFCR Clustering - Expectation-Maximization MSE vs MAE vs SC vs DB after PCA' ) print('HFCR Clustering - EPS (Density-based) after PCA') EPS = [2.5, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100] mse: list = [] mae: list = [] sc: list = [] db: list = [] rows, cols = ds.choose_grid(len(EPS)) _, axs = plt.subplots(rows, cols, figsize=(cols * 5, rows * 5), squeeze=False) i, j = 0, 0 for n in range(len(EPS)): estimator = DBSCAN(eps=EPS[n], min_samples=2) estimator.fit(data) labels = estimator.labels_ k = len(set(labels)) - (1 if -1 in labels else 0) if k > 1: centers = ds.compute_centroids(data, labels) mse.append(ds.compute_mse(data.values, labels, centers)) mae.append(ds.compute_mae(data.values, labels, centers)) sc.append(silhouette_score(data, labels)) db.append(davies_bouldin_score(data, labels)) ds.plot_clusters(data, eixo_x, eixo_y, labels.astype(float), estimator.components_, k, f'DBSCAN eps={EPS[n]} k={k}', ax=axs[i, j]) i, j = (i + 1, 0) if (n + 1) % cols == 0 else (i, j + 1) else: mse.append(0) mae.append(0) sc.append(0) db.append(0) plt.suptitle('HFCR Clustering - EPS (Density-based) after PCA') plt.savefig(subDir + 'HFCR Clustering - EPS (Density-based) after PCA') print( 'HFCR Clustering - EPS (Density-based) MSE vs MAE vs SC vs DB after PCA' ) fig, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.plot_line(EPS, mse, title='DBSCAN MSE', xlabel='eps', ylabel='MSE', ax=ax[0, 0]) ds.plot_line(EPS, mae, title='DBSCAN MAE', xlabel='eps', ylabel='MAE', ax=ax[0, 1]) ds.plot_line(EPS, sc, title='DBSCAN SC', xlabel='eps', ylabel='SC', ax=ax[0, 2], percentage=True) ds.plot_line(EPS, db, title='DBSCAN DB', xlabel='eps', ylabel='DB', ax=ax[0, 3]) plt.suptitle( 'HFCR Clustering - EPS (Density-based) MSE vs MAE vs SC vs DB after PCA' ) plt.savefig( subDir + 'HFCR Clustering - EPS (Density-based) MSE vs MAE vs SC vs DB after PCA' ) print('HFCR Clustering - Metric (Density-based) after PCA') METRICS = ['euclidean', 'cityblock', 'chebyshev', 'cosine', 'jaccard'] distances = [] for m in METRICS: dist = np.mean(np.mean(squareform(pdist(data.values, metric=m)))) distances.append(dist) print('AVG distances among records', distances) distances[0] = 80 distances[1] = 50 distances[2] = 80 distances[3] = 0.0005 distances[4] = 0.0009 print('CHOSEN EPS', distances) mse: list = [] mae: list = [] sc: list = [] db: list = [] rows, cols = ds.choose_grid(len(METRICS)) _, axs = plt.subplots(rows, cols, figsize=(cols * 5, rows * 5), squeeze=False) i, j = 0, 0 for n in range(len(METRICS)): estimator = DBSCAN(eps=distances[n], min_samples=2, metric=METRICS[n]) estimator.fit(data) labels = estimator.labels_ k = len(set(labels)) - (1 if -1 in labels else 0) if k > 1: centers = ds.compute_centroids(data, labels) mse.append(ds.compute_mse(data.values, labels, centers)) mae.append(ds.compute_mae(data.values, labels, centers)) sc.append(silhouette_score(data, labels)) db.append(davies_bouldin_score(data, labels)) ds.plot_clusters( data, eixo_x, eixo_y, labels.astype(float), estimator.components_, k, f'DBSCAN metric={METRICS[n]} eps={distances[n]:.2f} k={k}', ax=axs[i, j]) else: print(k) mse.append(0) mae.append(0) sc.append(0) db.append(0) i, j = (i + 1, 0) if (n + 1) % cols == 0 else (i, j + 1) plt.suptitle('HFCR Clustering - Metric (Density-based) after PCA') plt.savefig(subDir + 'HFCR Clustering - Metric (Density-based) after PCA') print( 'HFCR Clustering - Metric (Density-based) MSE vs MAE vs SC vs DB after PCA' ) fig, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.bar_chart(METRICS, mse, title='DBSCAN MSE', xlabel='metric', ylabel='MSE', ax=ax[0, 0]) ds.bar_chart(METRICS, mae, title='DBSCAN MAE', xlabel='metric', ylabel='MAE', ax=ax[0, 1]) ds.bar_chart(METRICS, sc, title='DBSCAN SC', xlabel='metric', ylabel='SC', ax=ax[0, 2], percentage=True) ds.bar_chart(METRICS, db, title='DBSCAN DB', xlabel='metric', ylabel='DB', ax=ax[0, 3]) plt.suptitle( 'HFCR Clustering - Metric (Density-based) MSE vs MAE vs SC vs DB after PCA' ) plt.savefig( subDir + 'HFCR Clustering - Metric (Density-based) MSE vs MAE vs SC vs DB after PCA' ) print('HFCR Clustering - Hierarchical after PCA') mse: list = [] mae: list = [] sc: list = [] db: list = [] rows, cols = ds.choose_grid(len(N_CLUSTERS)) _, axs = plt.subplots(rows, cols, figsize=(cols * 5, rows * 5), squeeze=False) i, j = 0, 0 for n in range(len(N_CLUSTERS)): k = N_CLUSTERS[n] estimator = AgglomerativeClustering(n_clusters=k) estimator.fit(data) labels = estimator.labels_ centers = ds.compute_centroids(data, labels) mse.append(ds.compute_mse(data.values, labels, centers)) mae.append(ds.compute_mae(data.values, labels, centers)) sc.append(silhouette_score(data, labels)) db.append(davies_bouldin_score(data, labels)) ds.plot_clusters(data, eixo_x, eixo_y, labels, centers, k, f'Hierarchical k={k}', ax=axs[i, j]) i, j = (i + 1, 0) if (n + 1) % cols == 0 else (i, j + 1) plt.suptitle('HFCR Clustering - Hierarchical after PCA') plt.savefig(subDir + 'HFCR Clustering - Hierarchical after PCA') print('HFCR Clustering - Hierarchical MSE vs MAE vs SC vs DB after PCA') fig, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.plot_line(N_CLUSTERS, mse, title='Hierarchical MSE', xlabel='k', ylabel='MSE', ax=ax[0, 0]) ds.plot_line(N_CLUSTERS, mae, title='Hierarchical MAE', xlabel='k', ylabel='MAE', ax=ax[0, 1]) ds.plot_line(N_CLUSTERS, sc, title='Hierarchical SC', xlabel='k', ylabel='SC', ax=ax[0, 2], percentage=True) ds.plot_line(N_CLUSTERS, db, title='Hierarchical DB', xlabel='k', ylabel='DB', ax=ax[0, 3]) plt.suptitle( 'HFCR Clustering - Hierarchical MSE vs MAE vs SC vs DB after PCA') plt.savefig( subDir + 'HFCR Clustering - Hierarchical MSE vs MAE vs SC vs DB after PCA') print('HFCR Clustering - Metric (Hierarchical) after PCA') METRICS = ['euclidean', 'cityblock', 'chebyshev', 'cosine', 'jaccard'] LINKS = ['complete', 'average'] k = 3 values_mse = {} values_mae = {} values_sc = {} values_db = {} rows = len(METRICS) cols = len(LINKS) _, axs = plt.subplots(rows, cols, figsize=(cols * 5, rows * 5), squeeze=False) for i in range(len(METRICS)): mse: list = [] mae: list = [] sc: list = [] db: list = [] m = METRICS[i] for j in range(len(LINKS)): link = LINKS[j] estimator = AgglomerativeClustering(n_clusters=k, linkage=link, affinity=m) estimator.fit(data) labels = estimator.labels_ centers = ds.compute_centroids(data, labels) mse.append(ds.compute_mse(data.values, labels, centers)) mae.append(ds.compute_mae(data.values, labels, centers)) sc.append(silhouette_score(data, labels)) db.append(davies_bouldin_score(data, labels)) ds.plot_clusters(data, eixo_x, eixo_y, labels, centers, k, f'Hierarchical k={k} metric={m} link={link}', ax=axs[i, j]) values_mse[m] = mse values_mae[m] = mae values_sc[m] = sc values_db[m] = db plt.suptitle('HFCR Clustering - Metric (Hierarchical) after PCA') plt.savefig(subDir + 'HFCR Clustering - Metric (Hierarchical) after PCA') print( 'HFCR Clustering - Metric (Hierarchical) MSE vs MAE vs SC vs DB after PCA' ) _, ax = plt.subplots(1, 4, figsize=(10, 3), squeeze=False) ds.multiple_bar_chart(LINKS, values_mse, title=f'Hierarchical MSE', xlabel='metric', ylabel='MSE', ax=ax[0, 0]) ds.multiple_bar_chart(LINKS, values_mae, title=f'Hierarchical MAE', xlabel='metric', ylabel='MAE', ax=ax[0, 1]) ds.multiple_bar_chart(LINKS, values_sc, title=f'Hierarchical SC', xlabel='metric', ylabel='SC', ax=ax[0, 2], percentage=True) ds.multiple_bar_chart(LINKS, values_db, title=f'Hierarchical DB', xlabel='metric', ylabel='DB', ax=ax[0, 3]) plt.suptitle( 'HFCR Clustering - Metric (Hierarchical) MSE vs MAE vs SC vs DB after PCA' ) plt.savefig( subDir + 'HFCR Clustering - Metric (Hierarchical) MSE vs MAE vs SC vs DB after PCA' ) plt.close("all") plt.clf()