Recommender
system is used widely in every field of life especially in e commerce, machine
learning, data mining etc In which it recommends user by using some history or
old data of user, the best example of the recommender system is AMAZON
recommender system it recommend user things on the basis of its previous buying
thing or experience. In this paper it is discussed that how recommender system
help us to predict student performance on the basis of its previous history or
record. This
technique is also used in technology enhanced learning for recommending
different resources, for example recommendation of papers, books or any other material.
Educational data mining is another important technique that has taken part
recently and in this paper it is discussed that how recommender system is used
in educational data mining to predict student performance using performance data.
For applying Recommender system for predicting student performance the data is
used from the Knowledge Discovery and Data Mining Challenge. Many research work
has been done in this field before for example association rule mining is use
to discover student performance data using IF-THEN rules and then generate
recommendation on the result base on these rules. Other proposed an equation
collaborative ?ltering for predicting student performance. Other techniques
used for predicting student performance is e.g  decision trees and Bayesian networks. Computer-aided tutoring systems is
also helpful for predicting performance. This
systems allow predictors to collect a huge  amount of information about students that how
they  interacts with the tutoring system
and also use past successes and  failure
of the student to predict its performance. The information of this system is
collected in a log which is named as “click-stream log. This log contain every
action of the student. The information contain in this log is given as an
equation below.

Quick-Stream
log Information:

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

“time, student, context, action”

“step no., student, context, actions,
duration, correct”

 :This equation is also used in quick-stream
log with additional information.

The two
most common recommender system techniques that  are used in this paper for predicting the student
performance are,

 

·        
Collaborative filtering

·        
Matrix Factorization

Collaborative
?ltering is the recommender technique which work on the basis of assumption means
its assume user interest on the basis of its previous history or record. Matrix
factorization technique is superior to classic nearest neighbor techniques for
recommendations of different things, In this techniques high correspondence
between students and hisher past performance leads to recommendation. We also
use root mean squared error (RMSE) this techniques is used for optimization of
result.

Two data
sets are used in the paper from Knowledge Discovery and Data Mining (KDD)
Challenge year 2010. Basically
data represents the log actions of interactions (given in above equation)
between students and computer-aided-tutoring systems. By this way doing same
process again and again they collect result of student performance during
computer-aided tutoring systems. After this they map all education data to both
Recommender system and regression problem. Both  linear
regression and logistic regression shows same result. We will not discuss in
detail here that how data will map on both techniques using educational data
set. Educational data
sets was map from the educational context to both recommender systems and
regression contexts. After this next step was to collaborative ?ltering and
regularized matrix factorization, for this implementation of algorithm was
implemented in My Media open source Framework (4).

Root
mean squared error (RMSE) is applied on different methods such as:

·        
Global Average

·        
User Average

·        
Logistic Regression

·        
Matrix Factorization

·        
User-Item Collaborative Filtering

·        
Matrix Factorization + Global Average

·        
Matrix Factorization + User Average

·        
Matrix Factorization + User-Item Collaborative
Filtering

·        
Matrix Factorization + User-Item Collaborative
Filtering

 

The best result was achieved by “Matrix
Factorization + User-Item Collaborative Filtering” using KC-RULES(knowledge
components Rules)