The aim of this project is to train an end-to-end model in order to predict the user audience's question using learning algorithms. It is highly recommended to take advantage of NLP approaches for better performance.
The Natural Language Processing Laboratory of the Faculty of Engineering and Computer Science of Shahid Beheshti University is developing a system called Soha (Intelligent Communication System) as the faculty chatbot and intelligent student assistant. In this way, while students gain the experience of participating in real practical work, they will be able to apply what they have learned during the semester.
Soha's task is to talk to the user, answer his questions, provide the necessary guidance or refer questions, criticisms, suggestions, or requests of the user to the relevant person responsible for answering. Therefore, in one of the small modules, it is necessary for the system to determine the addressee of a statement.
As we mentioned earlier, the purpose of this project to build a model to classify the user's query(question) into these 5 categories:
Category 1: Faculty Education aka. آموزش دانشکده
Example #1: آخرین مهلت حذف چه موقع است؟
Example #2: از کجا می تونم کارنامه مهر خورده بگیرم؟
Category 2: Information Desk aka. میز اطلاعات
Example #1: کتابخانه کجاست؟
Example #2: کجا می توانم رییس دانشکده را ببینم؟
Category 3: Site/Library aka. سایت/کتابخانه
Example #1: چگونه می توانم روي سرور دانشکده اکانت بگیرم؟
Example #2: اگر کتاب رو دیر بیارم چقدر جریمه میشم؟
Category 4: Information and Suggestions Box aka. صندوق اطلاعات و پیشنهادات
Example #1: نمیخواهید بالاخره کلاسها رو حضوري کنین؟
Example #2: آسانسور خراب شده، لطفا رسیدگی شود
Category 5: Others aka. سایر
Example #1: هوا خوب است یا هوا ابری است؟
Example #2: دیشب پرسپولیس چند تا از استقلال گل خورد؟
-
python
-
jupyterNotebook
-
HTML