modes = [ 'ind', 'iden', 'uni', 'fdiff', 'fmax', 'fsum', 'buc_eq', 'buc_con', 'buc_qeq', 'buc_qsd' ] W_name = np.array([ 'adult', 'age1', 'age2', 'age3', 'Total', 'Total', 'Identity', 'Prefix', 'Prefix' ]) W_lst = np.array([ census.__adult(), census.__age1(), census.__age2(), census.__age3(), workload.Total(n), workload.Total(n), workload.Identity(n), workload.Prefix(n), workload.Prefix(n) ]) A_lst = strategy_comp(W_lst, n, rep) results = [] names = [] total_errors = pd.DataFrame() mean_ratio_errors = pd.DataFrame() max_ratio_errors = pd.DataFrame() min_ratio_errors = pd.DataFrame() max_distances = pd.DataFrame() min_distances = pd.DataFrame() gini_coefficients = pd.DataFrame() mean_idenratio_errors = pd.DataFrame() max_idenratio_errors = pd.DataFrame()
x_mean = np.sum(x_data[:,0]*x_data[:,1])/np.sum(x_data[:,1]) mid = np.sum(x_data[:,1])/2 prefix = np.cumsum(x_data[:,1]) x_median = x_data[np.searchsorted(prefix, mid),0] x_mode = x_data[np.argmax(x_data[:,1]),0] x_per = [] for percentile in [0.1, 0.25, 0.75, 0.9]: loc = np.sum(x_data[:,1])*percentile x_per.append(x_data[np.searchsorted(prefix, loc),0]) meanW = np.vstack((x_data[:,0], np.ones(n))) meanW = matrix.EkteloMatrix(meanW) W_name = np.array(['Mean', 'Median', 'Mode', 'Per_10', 'Per_25', 'Per_75', 'Per_90']) W_lst = np.array([meanW, workload.Prefix(n), workload.Identity(n),workload.Prefix(n), workload.Prefix(n),workload.Prefix(n), workload.Prefix(n)]) A_lst = strategy_comp(W_lst, n, rep) q_lst = W_name ans_lst = np.array([x_mean, x_median, x_mode] + x_per) print(ans_lst) results = [] names = [] ks = [] total_errors = pd.DataFrame() max_ratio_errors = pd.DataFrame() inters = pd.DataFrame() for i in range(t): print(i, flush=True)
seed = args.seed t = args.t if seed is not None: np.random.seed(seed) print(experiment_name) print('n={}, k={}, eps={}, rep={}, seed={}, t={}'.format(n,k,eps,rep,seed,t)) conf = OrderedDict() conf['n']=n conf['k']=k conf['eps'] = eps conf['rep']=rep conf['seed'] =seed conf['t'] = t modes = ['ind', 'iden', 'uni', 'fdiff', 'fmax', 'fsum', 'buc_eq', 'buc_con', 'buc_qeq', 'buc_qsd'] W_name = np.array(['race1', 'race2', 'white', 'Total', 'Identity']) W_lst = np.array([census.__race1(), census.__race2(), census.__white(), workload.Total(n), workload.Identity(n)]) A_lst = strategy_comp(W_lst, n, rep) results = [] names = [] total_errors = pd.DataFrame() mean_ratio_errors = pd.DataFrame() max_ratio_errors = pd.DataFrame() min_ratio_errors = pd.DataFrame() max_distances = pd.DataFrame() min_distances = pd.DataFrame() gini_coefficients = pd.DataFrame() mean_idenratio_errors = pd.DataFrame() max_idenratio_errors = pd.DataFrame() min_idenratio_errors = pd.DataFrame() iden_gini_coefficients = pd.DataFrame() for i in range(t):
seed = args.seed t = args.t if seed is not None: np.random.seed(seed) print(experiment_name) print('n={}, k={}, eps={}, rep={}, seed={}, t={}'.format(n,k,eps,rep,seed,t)) conf = OrderedDict() conf['n']=n conf['k']=k conf['eps'] = eps conf['rep']=rep conf['seed'] =seed conf['t'] = t modes = ['ind', 'iden', 'uni', 'fdiff', 'fmax', 'fsum', 'buc_eq', 'buc_con', 'buc_qeq', 'buc_qsd'] W_name = np.array(['adult', 'age1', 'age2', 'age3', 'Total', 'Identity']) W_lst = np.array([census.__adult(), census.__age1(), census.__age2(), census.__age3(), workload.Total(n), workload.Identity(n)]) A_lst = strategy_comp(W_lst, n, rep) results = [] names = [] total_errors = pd.DataFrame() mean_ratio_errors = pd.DataFrame() max_ratio_errors = pd.DataFrame() min_ratio_errors = pd.DataFrame() max_distances = pd.DataFrame() min_distances = pd.DataFrame() gini_coefficients = pd.DataFrame() mean_idenratio_errors = pd.DataFrame() max_idenratio_errors = pd.DataFrame() min_idenratio_errors = pd.DataFrame() iden_gini_coefficients = pd.DataFrame() for i in range(t):