def compare_vec_tweet_strenght(tweet, row_number):
dist_list = []
vector = process_single_tweet(tweet)
##### Einlesen der Vektoren und berechnung der Cosinus-Distanz #####
with open('POS_Testdaten_vec_cb.csv', 'r') as f:
reader = csv.reader(f, delimiter=";")
for row in reader:
words = []
int_words = []
words = row[2].split(',') #Aufspalten der Vektoren
for word in words:
word = int(word)
int_words.append(word) #schreiben in Liste als Integer
cos = cos_sim(int_words, vector) #Berechnung der Distanz zwischen Test-Vektor und Vektor aus .csv
dist = {}
dist["CB"] = row[row_number]
dist["cos"] = cos
#dist[row[3]] = cos
dist_list.append(dist.copy())
sorted_dist = sorted(dist_list, key=lambda k: k['cos'], reverse=True)
knn_list = sorted_dist[:11]
cb = []
for x in knn_list:
cb.append(int(x.get("CB")))
#print(cb)
return median(cb)
def compare_vec_twitter_bullying_cb(tweet, row_number):
dist_list = []
vector = process_single_tweet(tweet)
##### Einlesen der Vektoren und berechnung der Cosinus-Distanz #####
with open('twitter_bullying_files/twitter_bullying_train_vec.csv', 'r') as f:
reader = csv.reader(f, delimiter=";")
for row in reader:
words = []
int_words = []
words = row[2].split(',') #Aufspalten der Vektoren
for word in words:
word = int(word)
int_words.append(word) #schreiben in Liste als Integer
cos = cos_sim(int_words, vector) #Berechnung der Distanz zwischen Test-Vektor und Vektor aus .csv
dist = {}
dist["CB"] = row[row_number]
dist["cos"] = cos
#dist[row[3]] = cos
dist_list.append(dist.copy())
sorted_dist = sorted(dist_list, key=lambda k: k['cos'], reverse=True)
knn_list = sorted_dist[:11]
#cb = []
cyberbullying = 0
for x in knn_list:
#print(x.get("CB"))
if x.get("CB") == "2":
cyberbullying = cyberbullying + 1
if x.get("CB") == "1":
cyberbullying = cyberbullying + 1
# for x in knn_list:
# cb.append(int(x.get("CB")))
# print(cb)
return round(cyberbullying/11, 2)
def compare_vec_bullying_traces(tweet, row_number):
dist_list = []
vector = process_single_tweet(tweet)
##### Einlesen der Vektoren und berechnung der Cosinus-Distanz #####
with open('bullying_traces_files/bullying_traces_train_vec.csv', 'r') as f:
reader = csv.reader(f, delimiter=";")
for row in reader:
words = []
int_words = []
words = row[2].split(',') #Aufspalten der Vektoren
for word in words:
word = int(word)
int_words.append(word) #schreiben in Liste als Integer
cos = cos_sim(
int_words, vector
) #Berechnung der Distanz zwischen Test-Vektor und Vektor aus .csv
dist = {}
dist["CB"] = row[row_number]
dist["cos"] = cos
#dist[row[3]] = cos
dist_list.append(dist.copy())
sorted_dist = sorted(dist_list, key=lambda k: k['cos'], reverse=True)
knn_list = sorted_dist[:11]
count_cb = 0
for x in knn_list:
cb = x.get("CB")
count_cb += int(cb)
#print(cb)
#percent += int(bc)
percent = count_cb / 11
return round(percent, 2)