As the development of information technology and Internet,
people enter the era of information overload from that of
information deficiency gradually[17]. The report is the result of
the Master Thesis in Information and Communication
Technology School at Royal Institute of Technology in
Stockholm, Sweden. The project is provided in VionLabs AB
that is a media-tech company providing media content such as
movie information for customers. The recommender system we
implemented is for the movie website.
During the past several decades personnel function has been
transformed from a relatively obscure record keeping staff to central
and top-level management function. important resources. This
project intends to introduce more user friendliness in the various
activities such all the details required for the correct statement
calculation and generation. as record updation, maintenance, and
searching. The searching of record has been made quite simple as all
the details of the customer can be obtained by simply keying in the
identification or account number of that customer. Similarly, record
maintenance and updation can also be accomplished by using the
account number with all the details being automatically generated.
These details are also being promptly automatically updated in the
master file thus keeping the record absolutely up-to-date. The entire
information has maintained in the database or Files and whoever wants to retrieve can't retrieve, only authorization user can retrieve
the necessary information which can be easily be accessible from the
file.
A computer-based recommendation system is designed to handle
all the primary information required to calculate user's emotions.
Separate database is maintained to handle all the details required
for the correct statement calculation and generation.
This project intends to introduce more user friendliness in the
various activities such as record updation, maintenance, and
searching. The searching of record has been made quite simple
as all the details of the customer can be obtained by simply
keying in the identification or account number of that customer.
Similarly, record maintenance and updation can also be
accomplished by using the account number with all the details
being automatically generated. These details are also being
promptly automatically updated in the master file thus keeping
the record absolutely up-to-date.
The main objective of our project is providing the different typed
of customers facility, the main objective of this system is to find
out the actual customer service. Etc. It should fulfill almost all
the process requirements of any recommendation system.
User-based and Item-based collaborative-filtering algorithms are all neighborhood-
based algorithm, there are also a lot of other collaborative-filtering algorithms. Hoff-
man raised Latent Class Model in this paper[13], the model connects user and item
by latent class, which considers that a user will.
2. Abstract:
● Recommender System is a toolhelping users find content and overcome
informationoverload. It predicts interests of users and makes recommendation
according to the interest model of users. Theoriginal content-based
recommender system is the continuation and development of collaborative
filtering, which doesn’t need the user’s evaluation for items. Instead, the
similarity is calculated based on the information of items that are chose by users,
and then make the recommendationaccordingly.
3. Problem Statement
● Aim: Build a movie recommendation system based on ‘MovieLens’ dataset.
● Wewish to integrate the aspects of personalization of user with the overall
features of movie such as genre, popularity etc.
● Solve a day to day problem of userswondering which movie to watch next.
4. Introduction:
● Recommendationsystems produces a ranked list of items on which a
usermight be interested, in the context of hiscurrent choice of an item.
● Wewish to integrate the aspects of personalization of user with the
overallfeatures of movie such as genre, popularity etc.
6. DataSet
● MovieLens review dataset (ml-latest-small)
○ Ratings: 100k
○ Movies: 9k
○ Users: 600
● Integrated the dataset with IMDB and TMDB data set publically available.
● Splitthe dataset into 80% training and 20% testing based on the User ID.
8. Models
1. Popularity based model
2. Content based model
3. CollaborativeFiltering
4. MatrixFactorization method
5. Combined model (SVD + CF)
6. Hybrid model
9. Collaborative Filtering and Content
Based Filtering:
1. CollaborativeFiltering: The recommendation system
recommends movies which are rated highly by similar users.
f(movies, user) ---> (movies)
2. Content Based: The recommendation system recommends
movies which are similar to selected movie.
f(movie) —> (movies)
12. Sample results when enough information of User
● User Id = 1
● User top genre list from User vector:
○ [‘Film-Noir’, ‘Animation’, ‘Musical’]:
13. Sample results when less information about user
● User Id = 9
● User top genre from User Vector:
○ [‘Fantasy’, ‘Western’, ‘Mystery’]
14. Takeaways:
1. Content based with genre isgood when a userhas less ratings.
2. Movie similaritymetric based on features like overview, taglines and
genre.
3. Item-item collaborative filtering works better than user-user
collaborative filtering.
4. KNNbased and SVD algorithms improve when global baselines are
added.
5. Combining the predictions and recommendations of different
models gives better results in terms of accuracy and quality of
recommendations.