1616# In[2]:
1717
1818
19- data = pd .read_csv ('naukri_com-jobs__20190701_20190830__30k_data.csv' )
19+ data = pd .read_csv ('naukri_com-jobs__20190701_20190830__30k_data.csv' )
2020
2121
2222# In[3]:
2828# In[4]:
2929
3030
31- user_profiles = data [['Uniq Id' , 'Role Category' , 'Location' , 'Job Experience Required' , 'Key Skills' ]]
32- job_postings = data [['Uniq Id' , 'Role Category' , 'Location' , 'Job Experience Required' , 'Key Skills' ]]
31+ user_profiles = data [['Uniq Id' , 'Role Category' ,
32+ 'Location' , 'Job Experience Required' , 'Key Skills' ]]
33+ job_postings = data [['Uniq Id' , 'Role Category' ,
34+ 'Location' , 'Job Experience Required' , 'Key Skills' ]]
3335
3436
3537# ### User Profile
4850
4951user_profiles_matrix = pd .get_dummies (user_profiles .drop ('Uniq Id' , axis = 1 ))
5052user_profiles_matrix = normalize (user_profiles_matrix ) # Normalize the matrix
51- similarity_matrix = cosine_similarity (user_profiles_matrix , user_profiles_matrix )
53+ similarity_matrix = cosine_similarity (
54+ user_profiles_matrix , user_profiles_matrix )
5255
5356
5457# ### Job recommendation
5962# Define the number of nearest neighbors to consider
6063k = 5
6164
65+ 6266def get_job_recommendations (user_id ):
6367 user_index = user_profiles [user_profiles ['Uniq Id' ] == user_id ].index [0 ]
64- similar_users = similarity_matrix [user_index ].argsort ()[::- 1 ][1 :k + 1 ] # Exclude the user itself
68+ similar_users = similarity_matrix [user_index ].argsort (
69+ )[::- 1 ][1 :k + 1 ] # Exclude the user itself
6570
6671 # Get job postings from similar users
6772 recommended_roles = []
6873 for user in similar_users :
6974 similar_user_id = user_profiles .iloc [user ]['Uniq Id' ]
70- similar_user_roles = data [data ['Uniq Id' ] == similar_user_id ]['Role Category' ].values
75+ similar_user_roles = data [data ['Uniq Id' ] ==
76+ similar_user_id ]['Role Category' ].values
7177 recommended_roles .extend (similar_user_roles )
7278
7379 # Filter out already interacted job roles
74- user_interacted_roles = data [data ['Uniq Id' ] == user_id ]['Role Category' ].values
75- recommended_roles = list (set (recommended_roles ) - set (user_interacted_roles ))
80+ user_interacted_roles = data [data ['Uniq Id' ]
81+ == user_id ]['Role Category' ].values
82+ recommended_roles = list (set (recommended_roles ) -
83+ set (user_interacted_roles ))
7684
7785 # Rank recommended roles based on frequency
78- recommended_roles = pd .Series (recommended_roles ).value_counts ().sort_values (ascending = False )
86+ recommended_roles = pd .Series (
87+ recommended_roles ).value_counts ().sort_values (ascending = False )
7988
8089 return recommended_roles .index .tolist ()
8190
91+ 8292# Example usage
8393user_id = '9be62c49a0b7ebe982a4af1edaa7bc5f'
8494recommended_roles = get_job_recommendations (user_id )
@@ -90,11 +100,4 @@ def get_job_recommendations(user_id):
90100# In[ ]:
91101
92102
93- 94- 95- 96103# In[ ]:
97- 98- 99- 100-
0 commit comments