| People always desire to find educational institutions that fulfill their requirements.It's very hard to know different features of institutions and especially from the perspective of students and their parents,who have real life experience with the institutions.Nowadays Institutions have their websites and many other online sources like Wikipedia has information.Social media and blogs where people share their experiences about these institution.But it is still hard to get all the information at one place.Second that data is not from students and their parent's perspective,so it can't be 100 % valid or true about institutions.However online different sources made it easy.Institutions websites,educational online repositories and people generated content have enough data to find institutions and their characteristics.I propose a solution for educational institutions evaluation,in which Data(student's reviews)will be collected from different sources.Using Stanford dependency parser and defined rules,aspects and sentiments expressed for these aspects of the institutions will be extracted from collected reviews,and will be clustered using Word Net Synonyms for five seed words for each category.Like for TEACHING(lecturer,professor,stud,class,course),Staff,Facilities,opportunities etc.then sentiments will be calculated for each aspects category using Afinn dictionary,sentiment analysis will be done,rating for each category will be calculated,and summarization will be performed using top 10 common most words for sentence creating using NLG.Output will also be drawn on graphs like ratings of all aspects categories via Bar chart,number and percentage of people mentioned each category will be drawn on Pie chart using Plotly.This will give evaluation of the institution,and will help to explain the institution in a short summary from different aspects a user always focus on. |
PDF Full Text Request