def generateFeatureDictionary(addresskeys, keyAddressCounts):
logger.info("generating feature list")
featurelist = defaultdict(list)
for keypressed in keyAddressCounts:
for key in addresskeys:
if key in keyAddressCounts[keypressed]:
featurelist[key].append(keyAddressCounts[keypressed][key])
else:
featurelist[key].append(0)
featurelist['class'].append(charName(keypressed.split('_')[0]))
#print featurelist['0x1f37']
logger.debug("featurelist length %d", len(featurelist))
return featurelist
def preprocess_outlier(self, data_numbers):
gb = self.df.groupby('class')
for value in data_numbers:
#print('ajaya',df['class'],'***********************',value)
#data = df['ground_truth'].loc[df['class'] == value]
#print(gb)
data = gb.get_group(value)
#print(data['ground_truth'])
rmo = outliers(data['ground_truth'])
temp_df = pd.DataFrame(rmo.remove_outliers())
#print("neupane",temp_df.head())
temp_df['class'] = charName(value)
#print('ajaya',temp_df['class'])
df_list.append(temp_df)
print(df.head())
df_all = pd.concat(df_list, axis=0)
return df_all
#training_guesses = list_of_guesses(train_distribution)
test_distribution = 'distribution/distribution_test.csv'
test_guesses = list_of_guesses(test_distribution)
#print(training_guesses)
#print(test_guesses)
file_out = open('results_modeled_with_confidence_avg.csv', 'w')
word_given = ""
for p_word in words:
p_word = p_word.lower()
print("***************************")
print(p_word)
keys = []
for a_char in p_word:
print(a_char)
keys.append(test_guesses[charName(a_char)]) #if you turn on append
print(keys)
x = [keys[i].index(charName(p_word[i])) for i in range(len(p_word))]
print(x)
ret = 0
#file_out = open('results_modeled_with_confidence_dfs.csv','w')
# for i in range(len(p_word)):
# z = 1
# for j in range(i+1,len(p_word)):
# z = z * len(keys[j])
# ret += x[i] * z
#file_out = open('results_modeled_with_confidence_pow.csv','w')
#ret = pow(max(x)+1,len(p_word))