Link to the Web Page: http://webpages.uncc.edu/~snagiset/cloudreport.html
Instructions to run the code:
Content based recommendation system: $ mkdir pagerank_classes javac -cp /usr/lib/hadoop/:/usr/lib/hadoop-mapreduce/ FinalContent.java -d pagerank_classes -Xlint jar -cvf recom.jar -C pagerank_classes/ . hadoop jar recom.jar org.myorg.FinalContent /user/vkundula/input1(movies.csv) /user/vkundula/input2(ratings.csv) The final output will be in collaboutput6
Collaborative based recommendation system: $ mkdir pagerank_classes javac -cp /usr/lib/hadoop/:/usr/lib/hadoop-mapreduce/ FinalCollaborative.java -d pagerank_classes -Xlint jar -cvf recom.jar -C pagerank_classes/ . hadoop jar recom.jar org.myorg.FinalCollaborative /user/vkundula/input(ratings.csv) The final output will be in contentoutput6
Hybrid: $ mkdir pagerank_classes javac -cp /usr/lib/hadoop/:/usr/lib/hadoop-mapreduce/ Hybrid.java -d pagerank_classes -Xlint jar -cvf recom.jar -C pagerank_classes/ . hadoop jar recom.jar org.myorg.Hybrid /user/vkundula/input1(contentoutput.txt) /user/vkundula/input2(collaborativeoutput.txt) /user/vkundula/output
Notes: All the output files have been provided with the submission All the other details have been provided in the webpage
Dataset: The dataset used is movielens dataset with over 1 million data items
Framework used: we used Hadoop MapReduce for the project
Fun thing about project: It is very interesting to know how we generally get the recommendations in idbm and netflix with the help of this project