University Admission

S ITY A D MI SS IO N
UNIV E R PREDICTION
INTRODUCTION
RELATED WORK
TYPES OF MACHINE LEARNING
STEPS INVOLVED
DATA VISUALIZATION
PREDICTIVE MODELING
LIMITATIONS
CONCLUSION
INTRODUCTIO
N
 A persons education plays a vital role in their life. While planning for education students
often have several questions regarding the courses, universities, job opportunities,
expenses involved, etc.
 The majority of international students studying in the USA are from India and China. In
the past decade
 It is seen that the number of students pursuing Masters in Computer Science field from
universities in the USA is quite high and we will be Focusing on those students.
 Majority of universities in the USA follow similar guidelines for providing admission to
students.
 Universities taking to consideration different factors like GRE,TOEFL,SOP and
undergraduate degree score.
 Based on the overall profile of the student decision is taken by the universities admission
team to admit or reject a particular candidate
Objectives of Research
 The primary objective is to solve the problems of students while applying for the
universities in abroad.
 We will be developing a university admission predictor system which will help the
students to predict the chances of their application being selected for a particular
university they wish to apply.
 Multiple machine learning algorithms were evaluated to develop the system
 Also we will be creating a simple user interface which will help the users to input the
data related to student profile and get thepredicted resultfor the application based on the
profile as output.
 This will thus eventually help students saving the extra amount of time and money they
have to spend at the education consultancy firms .
Related wORK
 There have been several project and studies performed on topics related to students admission into
universities. (Bibodi et al. (n.d.)) used multiple machine learning models
 Nave Bayes algorithm was used to predict the likelihood of success of an application, and multiple
classification algorithms like Decision Tree, Random Forest, Nave Bayes and SVM were compared and
evaluated based on their accuracy to select the best candidates for the college. Limitation of this research
as that it did only relied on the GRE, TOEFL and Undergraduate Score of the student and missed on taking
into consideration other important factors like SOP and LOR documents quality, past work experience,
technical papers of the students etc.
 Bayesian Networks were used by (Thi et al. (2007)) to create a decision support system for evaluating the
application submitted by international students in the university. The model was developed by applying data
mining techniques and knowledge discovery rules to the already existing in-house admission prediction
system of the university. (Mane (2016)) conducted a similar research that predicted the chance of a student
getting admission in college based on their Senior Secondary School. The performance of both the models
was good the only drawback was the problem statement was single university-centric.
 (Mishra and Sahoo (2016))conducted a research from a university point of view to predict
the likelihood of a student enrolling in the university after the have enquired about of
courses in the university. They used K-Means algorithm for clustering the students based
on different factors like feedback, family income, family occupation, parents qualification,
motivation etc. to predict if the student will enroll at the university or not. The objective of
the model was to increase the enrolment of the students in the university.
 GRADE system was developed by (Waters and Miikkulainen (2013)) to support the
admission process for the graduate students in the University of Texas Austin Department
of Computer Science. The main objective of the project was to develop a system that can
help the admission committee of the university to take better and faster decisions.
 The time required by the admission committee to review the applications was reduced by
74% but human intervention was required to make the final decision on status if the
application.
 Nadeshwar created a model using the Multiple Logistic regression algorithm, it was able
to achieve accuracy rate of 67% only
TYPES OF MACHINE
LEARNING
Supervised Learning
UnSupervised Learning and
Reinforcement Learning
STEPS INVOLVED
1.Define the problem
2.Generate your own hypothesis
3.Get the Dataset
4.Data Cleaning
5.Exploratory Data Analysis
6.Predictive Modeling
DATA VISUALIZATION
• We will visualize the data using matplotlib
• We will compare between different features of the data set to get to know which has highest effect on the targeted
variable
GRE Score Vs COA Bar graph TOEFL Score Vs COA Bar graph Research Vs COA heat map
CGPA VS Chance of Admit heat map Correlation heat map comparison between all features
IMPORTING LIBRARIES
We have to import all the required packages
will import a set of regressors to predict the required model and will split the data using train test spli
PREDICTIVE
MODELING
KNN Regression Model Ridge Regression Model

Linear Regression Model Random Forest Model
COMPARISON OF ACCURACY
BETWEEN DIFFERENT MODELS
LIMITATIONS
As discussed earlier in the limitation of the research we have created the
models based only on the data of Indian Students studying Masters in
Computer Science in the USA, we have considered only ten universities with
different rankings. In future, more data related to additional universities and
courses can be added to the system. Also, the system can be enhanced to a
web-based application by making changes to the Node-Red. Other
classification algorithms can be evaluated to resolve the problem if they
perform better than the current algorithm the system can be easily updated to
support the new algorithm by changing the server code in the Node-Red.
CONCLUSION
 The main objective of this research was to develop a prototype of the system that can be used by the
students aspiring to pursue their education in the USA. Multiple machine learning algorithms were
developed and used for this research. Linear Regression model compared with other models. The model
can be used by the students for evaluating their chances of getting shortlisted in a particular university with
an average accuracy of 79%.
 A simple user interface was developed to make the application interactive and easy to use for the users
from the non-technical background. Node-Red was used to create the user interface. The overall objective
of the research was achieved successfully as the system allow the students to save the extra amount of time
and money that they would spend on education consultants and application fees for the universities where
they have fewer chances of securing admission. Also, it will help the students to make better and faster
decision regarding application to the universities.
Presented by:
P.Rohith
M.Sahith
L.Nikhil Sai
M.Chandu
Nath
N.Nikhil
Chowdary

University Admission

Uploaded by

Copyright:

Available Formats

University Admission

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

University Admission

Uploaded by

Copyright:

Available Formats

S ITY A D MI SS IO N

KNN Regression Model Ridge Regression Model

You might also like