Professional Documents
Culture Documents
Movie Recommendation System Using Machine Learning With Python
Movie Recommendation System Using Machine Learning With Python
ON
Mrs. laxmiprasanna
(AUTONOMOUS)
CERTIFICATE
1
This is to certify that the Literature survey report entitled MOVIE RECOMMENDATION
SYSTEM USING MACHINE LEARNING WITH PYTHON, being submitted by Mr.
P.Chanderraju, bearing ROLL.NO:19K91A05F2, Mr. P.Varshith , bearing ROLL.NO:19K91A05F3,
Mr. K.Abhinavkrishna, bearing ROLL.NO:19K91A05B5, Mr. M.Pranavswaroop, bearing
ROLL.NO:19K91A05D4 in partial fulfilment of requirements for the award of the degree of Bachelor
of Technology in Computer Science and Engineering, to the TKR College of Engineering and
Technology is a record of bonafide work carried out by them under my guidance and supervision.
Name and Signature of the Guide Name and Signature of the HOD
Professor
CONTENTS
S.NO. Pg.no.
2
1 ABSTRACT
2 INTRODUCTION
3 LITERATURE SURVEY REPORT
3.1 Collaborative Filtering for Movie Recommendation using RapidMine
4 REFERENCES
1. ABSTRACT
3
A recommendation engine filters the information mistreatment totally different
algorithms and recommends the foremost relevant things to users. It 1st captures the
past behaviour of a client and supported that, recommends product that the users can be
seemingly to shop for. If a totally new user visits an e-commerce website, that website
won't have any past history of that user. Therefore, however will the positioning
approach advocating product to the user in such a scenario? One attainable answer
might be to recommend the popular product, i.e. the product that are high in demand.
Another attainable answer might be to advocate the product which might bring the most
profit to the business. 3 main approaches are used for our recommender systems. One is
Demographic Filtering i.e they provide generalized recommendations to each user,
supported picture show quality and/or genre. The System recommends identical movies
to users with similar demographic options. Since every user is totally different, this
approach is taken into account to be too straightforward. The basic plan behind this
technique is that movies that are a lot of common and critically acclaimed can have the
next likelihood of being likeable by the common audience. Second is content-based
filtering, wherever we have a tendency to try and profile the user’s interests
mistreatment data collected, and advocate things supported that profile. the opposite is
cooperative filtering, wherever we have a tendency to try and cluster similar users
along and use data regarding the cluster to create recommendations to the user.
2. INTRODUCTION
4
Recommender systems used in a various form of areas together with movies, music,
news, books, analysis articles, search queries, social tags, and merchandise normally.
Recommendation System is a filtration program whose prime goal is to predict the movie to a
user towards a domain-specific item. In our case, this domain-specific item is a movie, so the
most focus of our recommendation system is to filter and predict solely those movies that a
user would favor given some information concerning the user him or herself. There are many
alternative ways that to create movie recommendation system however we've selected the
content base recommender system in order that user will simply get the foremost similar
movies on the user's interest. As our recommender system recommends the top high five
movies as like movie that user is selected.
Related work:
Methodology:
6
and the Performance operators for every type. Apply Model Rating prediction
operator retrieve a trained model and a test as input. Output of the Apply Model
is practice to compute performance using the Performance operator. The
Performance operator compute the assessment of rating prediction error
methods: Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and
Normalized Mean Absolute Error (NMAE). These error measure values are
returned as a performance vector and an example set. Here using the data set
from Group Lens Research Project. Ratings are on a gauge of 1 to 5 (whole-star
ratings only). There are three files: ‘ratings.dat’,’users.dat’ and ‘movie.dat’. For
prediction of user ratings using collaborative filtering we requisite data with
three fields: userid, movieid and ratings. File 'ratings.dat‘ contains the user
ratings in the format: UserID, MovieID, Rating, Timestamp. Each user has at
least 20 ratings and surely ratings by users are sparse.
Conclusion:
7
2. Mining affective text to improve social media movies
recommendation
Related work:
Social media websites, such as YouTube and Flicker, are currently gaining in
popularity. A large volume of information is generated by online users and how
to appropriately provide personalized content is becoming more challenging.
Traditional recommendation models are overly dependent on preference ratings
and often suffer from the problem of “data sparsity”. Recent research has
attempted to integrate sentiment analysis results of online affective texts into
recommendation models; however, these studies are still limited. The one class
collaborative filtering (OCCF) method is more applicable in the social media
scenario yet it is insufficient for item recommendation. In this study, we
develop a novel sentiment-aware social media recommendation framework,
referred to as SA_OCCF, in order to tackle the above challenges. We leverage
inferred sentiment feedback information and OCCF models to improve
recommendation performance. We conduct comprehensive experiments on a
real social media web site to verify the effectiveness of the proposed framework
and methods. The results show that the proposed methods are effective in
improving the performance of the baseline OCCF methods.
Methodology:
Lately, social media websites (e.g., YouTube1 and Flicker2) are increasingly
receiving attention. A rapid convergence of online content sharing network
websites has been observed in recent years. A large volume of content can be
generated and diffused by users in these social media websites. For example, it
8
was reported in 2012 that YouTube received 60 h of uploads per minute and 4
billion views per day3. In such an environment, there is an urgent requirement
for an intelligent tool to effectively recommend social media items. Great
challenges on traditional recommendation techniques to provide personalized
content to users are seen, due to the dynamic behavior of users in social media
websites and the volume of content they generate. Traditional recommendation
models assume user preference ratings are available and often suffer from the
problem of “data sparsity”. In the real-world social media websites, it is hard to
obtain rating information which means that traditional models are limited in
their functionality. To address this issue, some researchers have explored rich
user generated content as a supplementary source to support personalized
recommendation. Tag-aware recommender systems were proposed to
incorporate rich tagging information into traditional recommendation models
and achieved good performance. Furthermore, a user’s online activity
(searching and browsing) and his/her social connections have also been
explored and utilized to improve recommendation accuracy. Recently, affective
texts (such as reviews and comments) generated by online users have been
given more attention and fruitful sentiment analysis works has begun to emerge
rapidly. Some attempts have been made to integrate sentiment analysis results
into recommendation generations. The majority of existing work combines
sentiment analysis techniques with collaborative filtering techniques to conduct
movie rating predication. However, in relation to the social media context, item
recommendation is limited due to a lack of sentiment-awareness. To the best of
the author’s knowledge, there are only two studies relevant to sentiment-aware
item recommendation in social media websites. These studies only applied
sentiment analysis results into the basic neighborhood-based collaborative
filtering models. Advanced one-class collaborative filtering (OCCF) models
were not explored. Therefore, the combination of sentiment analysis with a
recommender system is limited and more extensive experimental work is
9
required. In this paper, we develop a sentiment-aware social media
recommendation framework, referred to as SA_OCCF, to tackle the key
challenges highlighted above. Firstly, the affective text from user comments is
explored and mined by our proposed ensemble learning-based sentiment
classification (ELSC) method. Secondly, the derived sentiment feedback
information by ELSC is incorporated into the OCCF models and we formalize
our sentiment-aware recommendation models (SA_OCCF) to improve social
media item recommendation performance. The proposed methods and models
are evaluated through comprehensive experiments using the TED dataset. The
results show that the proposed SA_OCCF models outperform the baseline
methods using a variety of recommendation accuracy metrics.
Conclusion:
10
3. A Three Way Hybrid Movie Recommendation
Related work:
Methodology:
11
products or services, who has an opinion of that item. The input to
recommendation algorithm can be a database of user and items and output will
be the recommendations. As in this case, inputs consist of a dataset of customer
and database of movies and output denotes the movie recommendations.
algorithm [2] and review based text mining algorithm [3] . The content-based
algorithm consists of user’s information such as their interest, favorites,
priorities etc. This type of information cannot already be available for a
particular user; thus it is explicitly filled by the user. Content-based algorithms
provide recommendations or suggestions based on this type of inputs from the
user. Collaborative filtering is a technique that has been used in
recommendation systems which can predict and advise items that the user might
like based on his or her known preferences. CF-based movie recommendations
predict a list of top recommended movies for a given user based on ratings
(numerical scores) from many users & the available ratings that are explicitly
given by the user by rating a movie or movies [6] User ratings partially
represent user preferences. Thus, it is not easy to accurately identify similar
users based on the available ratings. On the other hand, online movie reviews
have become a common source for users to share and collect data about movies,
but there have been very few studies that accurately integrate text mining
techniques with traditional CF approaches to improve the suggestions based on
user’s preferences. Thus this paper proposes a combined approach of content
based algorithm, collaborative filtering and review based algorithm that
integrates user text reviews [3] and user numeric ratings [2] in order to model
user’s preferences better and in turn improve the performance of a movie
recommendation systems.
Conclusion:
12
User and Users having similar viewing preferences, by using text mining and
suggesting movies on the basis of intersection of the three algorithms i.e. User
Based Collaborative filtering, Content-based algorithm & text mining due to
which the user will not only be suggested movies but this system will also
provide the user with more refined recommendations as movies with a low
rating score in any of the Movie characteristics generated on the basis of Text
Mining will be filtered out during the priority allocation stage of the proposed
Three Way Hybrid Movie Recommendation System.
13
4. An intelligent movie recommendation system through group-level
sentiment analysis in microblogs
Related work:
The emergence of the online media sharing sites (e.g. Youtube, Youku, and
Hulu) have introduced new challenges in program recommendation in online
networks. However, there is a bottleneck that the amount of available viewing
logs and user friendship networks are too limited to design effective
recommendation algorithms. Thus, carrying out an intelligent program
recommendation system is important for these sites. In this work, we propose a
novel model which turns to the social networks and mine user preferences
information expressed in microblogs for evaluating the similarity between
online movies and TV episodes. To the best of our knowledge, it is the first
effort to bridge the gap between movie and TV watchers domain with social
network activities. Moreover, it is the first approach that can solve the “cold-
start” problem in movie and TV recommendation system. Series of data mining
approaches and social computing models have been adopted in this work.
Similar programs found from the social network are further used to suggest
programs in other media devices. This work can be easily applied in online
14
media streaming sites in order that intelligent recommendations of programs can
be made to the customers through mining microblogs.
Methodology:
In this paper, we turn to some other knowledge base such as Social Networks
(i.e., Twitter and Youtube) where people form different communities, each one
of which represents a group of fans for a specific program. Hence, it is possible
to identify and evaluate the communication and distance between such
communities. In this way, the relationship between different programs in social
network domain can be found, based on which we may further make
recommendations in movie/TV programs domain. For instance, we may suggest
John to watch Sports News during the break. KBridge explores the in-depth
connection among contents, thus can be used in the following usage scenarios:
• From a specific program, find other similar program according to the program-
wise association rules.
To the best of our knowledge, this is the first approach that matches the user
communication in social network domain with that in online movie streaming
system domain, and utilizes social network information and social computing
models in movie/TV program recommendation task.
15
patterns and deduce the association rule of movie/TV programs. Through this
way, we can overcome the second limitation, namely “cold-start” problem. To
the best of our knowledge, this is the first effort that bridges the gap between
user activity in program watching domain and social network domain.
• Moreover, our proposed model is generic in the aspect that, not only plain text,
but also other unstructured information within social networks can be utilized.
Conclusion:
16
5. Movie Recommendation System Using Collaborative Filtering
Related work:
Methodology:
17
utilities. These systems collect information about a user's preferences and
behavior, and then use this information to improve their suggestions in the
future. A large number of companies are making use of recommendation
systems to increase user interaction and enrich a user's shopping experience.
Recommendation systems have several benefits, the most important being
customer satisfaction and revenue. Many a times, customers tend to look at the
recommendations provided based on their previous transaction because they
think that they will find better options. If these recommendations are fine tuned
to the user's needs, the customer will be satisfied with their purchase. Thus, the
customer would use this application once again. With customers using these
applications frequently, a huge amount of revenue is generated, which is why
many e-commerce compames are turning to Improve their recommendation
engines. Although recommendation systems are common, developing systems
that provide good and appropriate suggestions is a challenge. Each user has
different preferences and likes. Additionally, a user's preference depends on
many aspects such as their mood, the occasion, the reason for their purchase etc.
If a website or app is not able to predict and provide suitable recommendations
as per the liking of the user, then the user is likely to stop using that website or
app. Thus, there is always a need for companies to improve their
recommendation systems. One goal of this paper IS to design a movie
recommendation system that considers the past movie ratings given by various
users to provide suggestions to the user. We implemented this system
collaborative filtering algorithms and Apache Mahout framework The second
goal is to compare the performance and efficiency of user-based recommender
system and item-based recommender system. This paper is organized as
follows: First, a brief overview of a few relevant, recent research done in the
space of recommender system will be discussed. Second, we will present the
understanding on the technique of collaborative filtering. Third, the data
18
preparation and data analysis approach using Mahout will be discussed. Finally,
a qualitative evaluation on the techniques used will be presented.
Conclusion:
19
6. Movie Recommendation System using Cosine Similarity and KNN
Related work:
Over the past years, the internet has broadened the horizon of various domains
to interact and share meaningful information. As it is said that everything has its
pros and cons therefore, along with the expansion of domain comes information
overload and difficulty in extraction of data. To overcome this problem the
recommendation system plays a vital role. It is used to enhance the user
experience by giving fast and coherent suggestions. This paper describes an
approach which offers generalized recommendations to every user, based on
movie popularity and/or genre. Content-Based Recommender System is
implemented using various deep learning approaches. This paper also gives an
insight into problems which are faced in content-based recommendation system
and we have made an effort to rectify them.
Methodology:
Advancement in technology is reaching new heights every day and due to which
we can see enormous growth in information. To deal with such large data we
20
use machine learning that automates analytical model building. The early
classification of machine learning is divided into three broad categories:
Supervised learning, Unsupervised learning and Reinforcement learning. We
use computers to make predictions to help us achieve better results using
various computational statistics. Tasks can be performed without being
explicitly programmed to do so. It becomes a tedious task to extract the relevant
information. Search engines solve the problem to some extent but it does not
solve the personalization problem. Recommendation System framework plays a
vital role in today‟s internet surfing, be it buying a product from an e-commerce
site or watching a movie on some video-on-demand service. In our everyday
life, we depend on recommendations given by other people either by word of
mouth or reviews of general surveys. People often use recommender systems
over the web to make decisions for the items related to their choice.
Recommendation systems are software tools and techniques whose goal is to
make useful and sensible recommendations to a collection of users for items or
products that might interest them. In other words, the recommender system or
recommendation systems belongs to a class of information filtering system that
aims at predicting the „preference‟ or „rating‟ given to an item. Revised
Manuscript Received on May 29.2020. Gaurav Srivastav, Inderprastha
Engineering College, AKTU Ramni Harbir Singh, Inderprastha Engineering
College, AKTU Sargam Maurya, Inderprastha Engineering College, AKTU
Tanisha Tripathi, Inderprastha Engineering College, AKTU Tushar Narula,
Inderprastha Engineering College, AKTU Recommendation systems are
primarily using three approaches. In content-based filtering, we do profiling
based on what type of content any user is interested in and using the collected
information, it recommends items. Another one is collaborative filtering, where
we make clusters of similar users and use that information to make
recommendations. Hybrid systems are the one which takes into account both
21
above stated approaches to deal with operational data more concisely. Our goal
is to provide accurate recommendations with less computational complexity.
Conclusion:
22
7. Content-Based Movie Recommendation System Using Genre
Correlation
Related work:
Methodology:
23
The quantity of data transactions has grown in this era of the Internet
exponentially. Massive numbers of users on the Internet have multiplied the
volume of data. Some data is accessible on the Internet, but not all of it is
meaningful or sufficient for use by the public. Excessive data turns out to be
inconsistent, and uncreative. Users must conduct several searches before they
eventually locate what they were searching for. Most importantly, they have
come up with recommendation systems to solve this problem. A
recommendation framework takes into account users' previous interests to
include relevant information. We strive to filter and configure the data
according to each user's requirements. Thanks to the growth of the Internet,
recommendation systems havebecome better and better That has gained
momentum because it provides details in a limited period of time.
Recommender programmes have been put in place in various media such as
books, videos, news, and goods, and they have proven effective there as well.
Recommender schemes are used for almost all organizations today. Indeed,
LinkedIn, Amazon, and Netflix are good examples of this trend. LinkedIn tells
the user which colleagues she may be connected to, among the millions who are
on the network, which ones they should subscribe to. By automating the
database, this means the user would not have to do a lot of manual research.
Since the Amazon recommendations are based on what consumers have already
bought together, they recommend products that go together. You can find
something that is new and noteworthy in a category you've selected on the home
page by following the recommendations from your favourites, as long as you're
still on the Amazon website. The same principle applies to Netflix as well; they
will determine what type of show a consumer would like and then recommend
similar shows. According to the ways of recommendation algorithms, three
methods can be used: content-based, collaborative, and hybrid. A behavioural
recommendation framework looks at a user's past behaviours and deduces
things about which they may be interested in. This type of search technique
24
involves taking previous user reviews and ratings and combines them with other
user profiles to form an assessment of what is trending or more common
Comparisons are rendered using those that have the most resemblance. On the
other hand, while content-based and collaborative filtering have their strengths,
there are also some important differences. Researchers came up with a solution
which would blend the strengths of both approaches. this paper proposes a
method that makes use of genre correlation This dataset was intended for this
reason, which has 9126 movies sorted into a genrebased classification scheme.
There are eleven genres in all. They have been reviewed by over a thousand six
hundred and seventy six-hundred users and seventy-one people. recommending
movies close to those that have earned good reviews by taking into account
Conclusion:
25
Security issues are key problem in mcc, they need to be focused more compare
to other issues.
Related work:
26
django framework. It preprocesses the data sets and build the algorithm on those
data sets. The data consists of thousands of movies of various genres. The
recommendation system takes one movie as input from the given data set and
outputs 5 movies as a recommendation result. A web based platform is built in
python programming language using django framework. These systems which
are based on content recommendation are mannered to people, these systems do
not recommend anything to the user, it limits to your choice.
Methodology:
27
fact that it turned out to be such. I get some stress. Please consider these Defects
recommendation system we developed Recommender system that recommends
movies to users It is based on the information you provide yourself. In In the
current study, users are their own Actor, director, Genre, year,rating, etc. User
choices are predicted based on: About selecting the history of previously visited
users. Or The system was developed in PHP and is now A simple console-based
interface.
Cosine Similarity:
Cosine similarity measures the similarity between two vectors in the inner
product space.It is measured by the sine and cosine of the angle between the
two vectors and determines if the two vectors point in approximately the same
direction.It is typically used in text analysis to measure document similarity.
Recommendation System:
The system takes a movie as input to recommend other similar movies based on
the cosine similarity machine learning algorithm.Recommendation System is a
web application that uses API call from The Movie Data base to fetch movie
library to get the recommended movie to user.
Conclusion:
28
possible. More meaningful results in our system. Besides us I want to integrate
various machine learning methods Examine the clustering algorithm and
comparison results. Finally, implement a web-based user Interface with user
database and learning model To suit each user.
Related work:
29
the user might be interested in. Content based filtering uses a series of distinct
and discrete characteristics of an item in order to recommend more items with
same properties. Both of these systems combine to make a hybrid recommender
system. This system which is a hybrid of both filtering systems is capable of
recommending movies using analysis of the profiles.
Methodology:
Conclusion:
Published by: Manoj Kumar, D.K. Yadav, Ankur Singh and Vijay Kr. Gupta
Related work:
Now a day’s recommendation system has changed the style of searching the
things of our interest. This is information filtering approach that is used to
predict the preference of that user. The most popular areas where
recommender system is applied are books, news, articles, music, videos,
movies etc. In this paper we have proposed a movie recommendation
system named MOVREC. It is based on collaborative filtering approach
that makes use of the information provided by users, analyzes them and then
recommends the movies that is best suited to the user at that time. The
recommended movie list is sorted according to the ratings given to these
movies by previous users and it uses K-means algorithm for this purpose.
MOVREC also help users to find the movies of their choices based on the
movie experience of other users in efficient and effective manner without
31
wasting much time in useless browsing. This system has been developed
in PHP using Dreamweaver 6.0 and Apache Server 2.0. The presented
recommender system generates recommendations using various types of
knowledge and data about users, the available items, and previous transactions
stored in customized databases. The user can then browse the
recommendations easily and find a movie of their choice.
Methodology:
In today’s world where internet has become an important part of human life,
users often face the problem of too much choice. Right from looking for a
motel to looking for good investment options, there is too much information
available. To help the users cope with this information explosion,
companies have deployed recommendation systems to guide their users. The
research in the area of recommendation systems has been going on for
several decades now, but the interest still remains high because of the
abundance of practical applications and the problem rich domain. A number of
such online recommendation systems implemented and used are the
recommendation system for books at Amazon.com , for movies at
MovieLens.org, CDs at CDNow.com (from Amazon.com), etc. Recommender
Systems have added to the economy of the some of the e-commerce
websites (like Amazon.com) and Netflix which have made these systems a
salient parts of their websites.
32
biggest e-commerce websites (like Amazon.com, snapdeal.com) and the online
movie rental company Netflix to make these systems a salient part of their
websites. High quality personalized recommendations add another
dimension to user experience. The web personalized recommendation systems
are recently applied to provide different types of customized information to
their respective users. These systems can be applied in various types of
applications and are very common now a day.
Conclusion:
33
11.Machine Learning Model for Movie Recommendation System
Related work:
34
have to increase our CineMatch algorithm 10% by using fashionable
collaborative filtering techniques.
Methodology:
In the User-Item matrix, each row represents a person and every column
represents an object and every cell represents rating given with the id of a user
to an item.
Here, two customers could be similar to the premise of the comparable ratings
given with the id of each of them. If any two users are similar then it means
both of them have given very comparable scores to the items due to the fact
here the consumer vector is nothing however the row of a matrix which in flip
contains rankings given through user to the items. Now considering cosine
similarity can variety from ‘0’ to ‘1’ and ‘1’ means the highest similarity, so
consequently, all the diagonal elements could be ‘1’ because the similarity of
the consumer with him/herself is the highest. But there's one hassle with user-
user similarity. User alternatives and tastes change over time. If any consumer
favored some item one year in the past then it isn't important that he/she will
like the identical object even today.
Here, two items can be comparable to the idea of the comparable rankings given
to each of the items via all of the users. If any two gadgets are comparable then
it means both of them had been given very comparable ratings by means of all
of the users due to the fact here the item vector is nothing however the column
of the matrix which in flip contains scores given with the aid of consumer to the
35
objects. Now due to the fact cosine similarity can variety from ‘0’ to ‘1’ and ‘1’
means the highest similarity, so consequently, all of the diagonal elements
might be ‘1’ due to the fact the similarity of an item with the identical item is
the highest.
The cold start problem concerns the personalized guidelines for users without a
few past histories (new users). Providing suggestions to users with small beyond
history turns into tough trouble for CF models due to the fact their studying and
predictive ability is limited.
Conclusion:
So, far our best model is SVDpp with Test RMSE of 1.0675.Here we are not
much worried about our RMSE because we haven’t trained it on the whole
data . Our main intention here is to learn more about Recommendation
Systems .If we taken whole data we would definitely get better RMSE . Tune
hyper parameters of all the Xgboost models above to improve the RMSE. Here
we used 10K users and 1K movies to train the above models due to my pc ram
issues. In the future, I am going to run on the entire information set using cloud
resources.
36
12.Content Based Movie Recommendation System
Related work:
37
recommendation system describes the items that may be recommended to the
user. Based on a data set, it predicts what movies a user will like considering the
attributes present in the previously liked movies. Recommendation systems can
recommend movies based on one or a combination of two or more attributes.
While designing a movie recommendation system various factors are
considered such as the genre of the movie, the director or the actors present in
it. In this paper, the recommendation system has been built on cast, keywords,
crew, and genres. A single column is created which will be the sum of all the 4
attributes, and it acts as a dominant factor for this movie recommender system.
Methodology:
The enhancement of science and technology leads to make the life more
comfortable than older days. The emerging technologies like neutrosophic
shortest path, transportation problem, uncertainty problem, fuzzy shortest path,
powershell, wireless sensor network, computer language, neural network,
routing, image processing making the products more intelligent and self-healing
based. The smart city applications like smart water, smart grid, smart parking,
smart resource management, etc. are based on IoT and IoE technologies. In this
manuscript, the recommendation system has been built on cast, keywords, crew,
& genres. The recommendation system aims to predict or take users’ interests
and recommend related items that quite likely are interesting for them. The
growth in the amount of information that is available online and the increase in
the number of Internet users has created an overload of information which
makes it difficult to find the correct information at the right time. The
recommender system solves this problem by filtering the required data from a
large amount of information that is generated based on the user’s interest or
preferences. Recommender systems are used for recommending products,
generating playlists, matchmaking, and a lot more. Recommender systems
function with characteristic information and user-item interactions.
38
Characteristic information is the information about the user and the items
whereas user-item interaction is the information regarding ratings, the number
of purchases, likes of the users, and many more. Based on this, the
recommendation system can be developed using collaborative filtering, content-
based filtering, or hybrid filtering.
Collaborative Filtering:
This system identifies users with similar tastes and uses their opinion to
recommend the same to another user with similar interest. It generates
recommendations using information about rating profiles for different users or
items. It has been implemented in different applications such as YouTube,
Netflix, and Spotify. It is a widely used approach and is used as a part of the
hybrid system.
Content-Based Filtering:
Hybrid Approach:
39
approach and later combining them. It increases the accuracy and performance
of the recommender systems..
Conclusion:
Related work:
40
Today's online world was fully filled up with blogs, views, comments, posts
through various websites and social-surfs. People were habituated with posting
every incident into blogs, messed with comments like text and emotions, which
are a mixed bag of sad, happy, worry, cry etc. Analysing such data was called as
Sentimental Analysis. To analysis, these unordered data we use new emerged
technology algorithms. Machine learning a transpire technology which is
engaged with almost all the fields, where its algorithms are more powerful that
give with better faultless results. In this paper, we are analyzing tweets based on
movie reviews using the Multinomial Logistic Regression, Naïve Bayes, and
SVM algorithms to compare score value to show the best text analysis
algorithm.
Methodology:
A model is proposed to analyze the data using tweet dataset. There are few steps
that process the tweets
b. Pre-processing
c. Creating dataset
d. Applying Algorithms
Firstly, to get data for sentiment, we need to retrieve data from Twitter. We
need to access them through Twitter API. Steps as follows,
41
• After the application is created with your twitter account, we need to access
the Access Tokens and Keys, they provided. consumer_secret: < Your Twitter
Consumer Secret Key> consumer_key: <Your Twitter Consumer Key>
access_token: < Your Twitter Access Token Key > access_secret: < Your
Twitter Access Secret Key >
Pre-processing of datasets:
• Substitute all emotag with their respective sentiments like happy sad, fear etc.
Creating Datasets:
To process data and extract the file as .csv into your directory,
• For data, pre-processing, extract the data from a text file to process it.
• Finally, it automatically creates the .csv file into the same directory.
42
Conclusion:
Sentiment Analysis is just a part of social media monitoring platform which will
be beneficial for a business to improve the productivity. It shows the ones
feeling or opinion over an object. We conclude by that Multinomial Naïve
Bayes with Machine Learning algorithm that produces an improvised result
when compared to the other classifier algorithms for language processing like
Multinomial Logistic Regression and Support Vector Machine (SVM).
Published by: Zan Wang, Xue Yu* , Nan Feng, Zhenhua Wang
Related work:
43
Recommendation systems have become prevalent in recent years as they
dealing with the information overload problem by suggesting users the most
relevant products from a massive amount of data. For media product, online
collaborative movie recommendations make attempts to assist users to access
their preferred movies by capturing precisely similar neighbors among users or
movies from their historical common ratings. However, due to the data sparsely,
neighbor selecting is getting more difficult with the fast increasing of movies
and users. In this paper, a hybrid model-based movie recommendation system
which utilizes the improved K-means clustering coupled with genetic
algorithms (GA) to partition transformed user space is proposed. It employs
principal component analysis (PCA) data reduction technique to dense the
movie population space which could reduce the computation complexity in
intelligent movie recommendation as well. The experiment results on Movielens
dataset indicate that the proposed approach can provide high performance in
terms of accuracy, and generate more reliable and personalized movie
recommendations when compared with the existing methods.
Methodology:
44
recommendation algorithms and extensions. The majority of existing
recommendation systems is based on collaborative filtering (CF) mechanism
which has been successfully developed in the past few years. It first collects
ratings of movies given by individuals and then recommends promising movies
to target customer based on the “like-minded” individuals with similar tastes
and preferences in the past. There have been many famous online multimedia
platforms (e.g., youtube.com, Netflix.com, and douban.com) incorporated with
CF technique to suggest media products to their customers. However, traditional
recommendation systems always suffer from some inherent limitations: poor
scalability, data sparsity and cold start problems. A number of works have
developed model-based approaches to deal with these problems and proved the
benefits on prediction accuracy in RS. Model-based CF uses the user-item
ratings to learn a model which is then used to generate online prediction.
Clustering and dimensionality reduction techniques are often employed in
model-based approaches to address the data sparse problem. The sparsity issues
arise due to the insufficiency of user’s history rating data and it is made even
more severe in terms of the dramatically growth of users and items. Moreover,
high-dimensional rating data may cause it difficult to extract common
interesting users by similarity computation, which results in poor
recommendations. In the literature, there have been many model-based
recommendation systems developed by partitioning algorithms coupled, such as
Kmeans and self-organizing maps (SOM). The aim of clustering is to divide
users into different groups to form “like-minded” (nearest) neighbors instead of
searching the whole user space, which could dramatically improve the system
scalability. It has been proved that clustering-based recommendation systems
outperform the pure CF-based ones in terms of efficiency and prediction
quality. In many works, the clustering methods are conducted with the entire
dimensions of data which might lead to somewhat inaccuracy and consume
more computation time. In general, making high quality movie
45
recommendations is still a challenge, and exploring an appropriate and
efficiency clustering method is a crucial problem in this situation.
Conclusion:
46
15.Matrix Factorization Techniques for Recommender Systems
Related work:
47
As the Netflix Prize competition has demonstrated, matrix factorization models
are superior to classic nearest neighbor techniques for producing product
recommendations, allowing the incorporation of additional information such as
implicit feedback, temporal effects, and confidence levels.
Modern consumers are inundated with choices. Electronic retailers and content
providers offer a huge selection of products, with unprecedented opportunities
to meet a variety of special needs and tastes. Matching consumers with the most
appropriate products is key to enhancing user satisfaction and loyalty.
Therefore, more retailers have become interested in recommender systems,
which analyze patterns of user interest in products to provide personalized
recommendations that suit a user’s taste. Because good personalized
recommendations can add another dimension to the user experience, e-
commerce leaders like Amazon.com and Netflix have made recommender
systems a salient part of their websites. Such systems are particularly useful for
entertainment products such as movies, music, and TV shows. Many customers
will view the same movie, and each customer is likely to view numerous
different movies. Customers have proven willing to indicate their level of
satisfaction with particular movies, so a huge volume of data is available about
which movies appeal to which customers. Companies can analyze this data to
recommend movies to particular customers.
Methodology:
Some of the most successful realizations of latent factor models are based on
matrix factorization. In its basic form, matrix factorization characterizes both
items and users by vectors of factors inferred from item rating patterns. High
correspondence between item and user factors leads to a recommendation.
These methods have become popular in recent years by combining good
48
scalability with predictive accuracy. In addition, they offer much flexibility for
modeling various real-life situations. Recommender systems rely on different
types of input data, which are often placed in a matrix with one dimension
representing users and the other dimension representing items of interest. The
most convenient data is high-quality explicit feedback, which includes explicit
input by users regarding their interest in products. For example, Netflix collects
star ratings for movies, and TiVo users indicate their preferences for TV shows
by pressing thumbs-up and thumbs-down buttons. We refer to explicit user
feedback as ratings. Usually, explicit feedback comprises a sparse matrix, since
any single user is likely to have rated only a small percentage of possible items.
One strength of matrix factorization is that it allows incorporation of additional
information. When explicit feedback is not available, recommender systems can
infer user preferences using implicit feedback, which indirectly reflects opinion
by observing user behavior, including purchase history, browsing history,
search patterns, or even mouse movements. Implicit feedback usually denotes
the presence or absence of an event, so it is typically represented by a densely
filled matrix.
Conclusion:
49
REFERENCES:
[1] Jain, A., & Vishwakarma, S. K., Collaborative Filtering for Movie
Recommendation using RapidMiner International Journal of Computer
Applications (0975 - 8887) Vol. 169 - No. 6, 2017.
50
[2] J. Sun, G. Wang, X. Cheng, and Y. Fu, Mining affective text to improve
social media movies recommendation, Information Processing & Management,
vol. 51, no. 4, pp. 444–457, 2015.
[3] K. Soni, R. Goyal, B. Vadera, and S. More, A Three Way Hybrid Movie
Recommendation System, International Journal of Computer Applications, vol.
160, no. 9, pp. 29–32, 2017.
[5] Wu, C.-S. M., Garg, D., & Bhandary, U. (2018). Movie Recommendation
System Using Collaborative Filtering. 2018 IEEE 9th International Conference
on Software Engineering and Service Science (ICSESS).
doi:10.1109/icsess.2018.8663822.
[7] Reddy, S., Nalluri, S., Kunisetti, S., Ashok, S., & Venkatesh, B. (2018).
Content-Based Movie Recommendation System Using Genre Correlation.
Smart Innovation, Systems and Technologies, 391–397. doi:10.1007/978-981-
13-1927-3_42.
51
[9] Movie Recommendation System Ananya Agarwal1, S. Srinivasan I
[10] Manoj Kumar, D.K. Yadav, Ankur Singh and Vijay Kr. Gupta (2015), A
Movie Recommender System: MOVREC, International Journal of Computer
Applications (0975 – 8887) vol. 124 – No. 3.
[13] Uma Ramya, V., & Thirupathi Rao, K. (2018). Sentiment Analysis of
Movie Review using Machine Learning Techniques. International Journal of
Engineering & Technology, 7(2.7), 676. doi:10.14419/ijet.v7i2.7.10921.
[14] Wang, Z., Yu, X., Feng, N., & Wang, Z. (2014). An improved
collaborative movie recommendation system using computational intelligence.
Journal of Visual Languages & Computing, 25(6), 667–675.
doi:10.1016/j.jvlc.2014.09.011.
52