Spaces:
Sleeping
Sleeping
Vipul-Chauhan
commited on
Commit
•
55e37de
1
Parent(s):
1e35828
Update app.py
Browse files
app.py
CHANGED
@@ -137,14 +137,18 @@ def cosine_similarity(a,b):
|
|
137 |
|
138 |
def return_selected_cluster(ques):
|
139 |
ques_clean = process_row(ques)
|
|
|
140 |
cluster_selected =-1
|
141 |
cluster_score =0
|
142 |
for clus_id in dictt_cluster_words:
|
143 |
score_temp=0
|
|
|
144 |
for word in ques_clean.split():
|
145 |
dictt_temp = dictt_cluster_words[clus_id]
|
146 |
if word in dictt_temp:
|
|
|
147 |
score_temp+=dictt_temp[word]
|
|
|
148 |
if score_temp>cluster_score:
|
149 |
cluster_selected = clus_id
|
150 |
cluster_score = score_temp
|
|
|
137 |
|
138 |
def return_selected_cluster(ques):
|
139 |
ques_clean = process_row(ques)
|
140 |
+
count_tokens = len(ques_clean.split())
|
141 |
cluster_selected =-1
|
142 |
cluster_score =0
|
143 |
for clus_id in dictt_cluster_words:
|
144 |
score_temp=0
|
145 |
+
matched_token=0
|
146 |
for word in ques_clean.split():
|
147 |
dictt_temp = dictt_cluster_words[clus_id]
|
148 |
if word in dictt_temp:
|
149 |
+
matched_token+=1
|
150 |
score_temp+=dictt_temp[word]
|
151 |
+
score_temp*= (matched_token/count_tokens)
|
152 |
if score_temp>cluster_score:
|
153 |
cluster_selected = clus_id
|
154 |
cluster_score = score_temp
|