print(s)
arq_destino.write(s + '\n')
s = ' - f_statistic = %.15f / p_value = %.15f' % (f_statistic, p_value)
print(s)
arq_destino.write(s+'\n')
colunas_desc = udata.COL_NAMES
for prob in udata.PROBABILIDADES:
for size in udata.TAMANHOS:
df = udata.obterDados2()
df = udata.filtrarPorProbabilidadeErro(df, prob=prob)
df = udata.filtrarPorTamanhoArray(df, tamanho=size)
if (df.shape[0] > 0):
arq_name = 'anova_prob_%s_tam_%s' % (prob, size)
arq_destino = os.path.join(path_arq_destino, '%s.txt' % (arq_name))
if os.path.exists(arq_destino):
os.remove(arq_destino)
arq_destino = open(arq_destino, 'w+')
csv_destino = os.path.join(path_arq_destino, 'csv/%s' % (arq_name))
head_n = 1000
df_bubble = df[df[udata.obterNomeColuna('algoritmo')] == 'bubble']#.head(head_n)
df_merge = df[df[udata.obterNomeColuna('algoritmo')] == 'merge']#.head(head_n)
import numpy as np
import pandas as pd
import seaborn as sns
plt.style.use('seaborn-whitegrid')
import util_graficos as graf
import util_ler_dados as udata
from scipy.stats.distributions import norm
from scipy.stats import kde
df = udata.obterDados()
# for prob in udata.PROBABILIDADES:
df_prob = udata.filtrarPorProbabilidadeErro(df, '0.01')
print(df_prob.shape)
print(df_prob['algoritmo'].unique())
# print( df_prob.groupby(['size_of_array', 'algoritmo']).count().head() )
sns.lmplot(
x='percentual_k_unordered',
hue='algoritmo',
y='percentual_maior_array',
data=df,
# markers=['o','v','+','x'],
scatter_kws={'s': 5},
col='algoritmo',
row='probabilidade_erro', # 'size_of_array',
height=3,
