Scribd
Upload
131 views15 pages

Movi3 Recommender System

This document discusses neural networks and fuzzy logic in recommender systems. It provides an overview of recommender systems, how they work, parameters considered, and algorithms used including content-based filtering and collaborative filtering. It also compares the results of a movie recommendation project to a research paper, showing improved accuracy of around 4.06% on the same MovieLens 100k dataset.

Uploaded by

Dhananjay Jahagirdar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
131 views15 pages

Movi3 Recommender System

This document discusses neural networks and fuzzy logic in recommender systems. It provides an overview of recommender systems, how they work, parameters considered, and algorithms used including content-based filtering and collaborative filtering. It also compares the results of a movie recommendation project to a research paper, showing improved accuracy of around 4.06% on the same MovieLens 100k dataset.

Uploaded by

Dhananjay Jahagirdar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 15

Neural Networks and Fuzzy Logic – CS 620

Anoop Krishna Y   01JST17CS022


Dhananjay V J       01JST17CS046
Santosh S                01JST17CS140
Why is Recommender System important ?
 Recommender System is beneficial to both service
providers and users
What is a Recommender System ?

People

Ads Movies

Product Recommender
Songs
s System
Movie Recommendation

Movie Watched
Movie Recommendation

Movie Watched Movie Recommended


Parameters considered for recommendation

• User Ratings
• No. of Reviews
• Relevancy

YouTube Considers 
• High Watch Time
• Engagement
• Topic Relevancy
• Video History
Classification of the possible result of a recommendation. 

Recommended   Not Recommended

True-Positive (tp)
Used False-Negative (fn)

Not Used False-Positive (fp) True-Negative (tn)


Item Similarity 

To build the set of expected movies, the system calculates the distance  between two movies by
measuring the relevance of these movies. In our experiments, we use collaborative-based similarity for
the item distance. 

Pearson correlation coefficient


Research Papers Referred 
•  Roy, Sushmita, Mahendra Sharma, and  Santosh Kumar  Singh. "Movie Recommendation System Using
Semi-Supervised Learning." In 2019 Global Conference for Advancement in Technology (GCAT), pp. 1-
5. IEEE, 2019.

• M. Robillard, R. Walker, and T. Zimmermann. Recommendation systems for software engineering.


IEEE Software, 27(4):80–86, July 2010.

Evaluation
• Shani, G., & Gunawardana, A. Evaluating recommendation systems, In Recommender Systems
Handbook, SpringerVerlag, New York, NY, USA, (2011), pp. 257-297.

Dataset
• Harper,  F.  Maxwell,  and  Joseph  A.  Konstan.  "The  movielens  datasets:  History  and  context." Acm
transactions on interactive intelligent systems (tiis) 5. In 2018  IEEE  International  Conference  on 
Intelligent Computer Communication and Processing (ICCP), pp. 93-99. IEEE, 2018
Comparing with the Research paper

Neighborhood
Our Project Research paper 1
size
10 0.863767 0.672917

20 0.885014 0.689355

30 0.905520 0.709531

40 0.920979 0.717344

Dataset used : MovieLens 100K movie ratings. 100,000 ratings from 1000 users on 1700 movies
Both were evaluated on the same 100k dataset.
The improvement observed is about ~4.06%.
 Basic Algorithms used in Recommender
Systems

• Content Base Filtering


• Collaborative Filtering                        
      
Collaborative Filtering
 Ex: Flipkart, Amazon                                                                                                                     

You might also like