A Review Study OF Movie Recommendation Using Machine Learning

International Research Journal of Engineering and Technology (IRJET) e-ISSN:2395-0056

Volume: 09 Issue: 11 | Nov 2022 www.irjet.net p-ISSN:2395-0072

A Review Study OF Movie Recommendation Using Machine Learning

Devani Radhika1, Saket Swarndeep2

1Student of Masters of Engineering, Ahmedabad, Dept. of Computer Engineering, L.J. University, Gujarat, India 2Assistant Professor, Ahmedabad, Dept. of Computer Engineering & Technology, Gujarat, India ***

Abstract - Recommendation systems forecast consumer preferences for yet-to-be-viewed goods like movies, songs, or books. The goal of movie recommendation systems is to anticipate users' interests and suggest products that are probably interesting to them. To assist users in finding the mostpertinentfilms,movierecommendationsystemsemploya variety of filtration techniques and algorithms .From a commercialstandpoint,themorerelevantinformationorfilms a user discovers on a certain platform, the better their engagement and, thus, the higher revenue. Numerous platformshavealsoshownthatsolelyreferralsaccountfor35 to 40% of revenue. Content-based filtering, collaborative filtering systems, and hybrid filtering are the most widely utilized subcategories of machine learning algorithms for movie suggestions. We can create a model that can propose movies based on historical data using machine learning. In this essay, we will examine various machine learning techniques that can be appliedto movie recommendation.

Key Words: Movie recommendation system, collaborative filtering, content-based filtering, hybrid system, k-means, k-nearest neighbour (KNN), clustering algorithm

1 INTRODUCTION

Theconceptofmachinelearninghademergedasearlyasthe late 19th century [1]. One of the main reasons we need a recommendersystemisthatinmoderntimes,withthehelp oftheinternet,wehavefartoomanyoptionsforeverything! Becauseoftherevolutionintheentertainmentindustry,the sourceofentertainmenthasgrownrapidlyintoday'sworld. Arecommendationsystemorrecommendationengineisa modelusedforinformationfilteringwhereittriestopredict thepreferencesofuserandprovidesuggestsbasedonthese preferences[2].Therecommendationsystemismostlyused indigitalentertainmentplatformslikeNetflix,PrimeVideo, and IMDB, as well as e-commerce platforms like Amazon, Flipkart,andeBay.

MoviesuggestionsforusersdependonWeb-basedportals. Movies can be easily differentiated through their genres, such as comedy, thriller, animation, and action. Another possiblewaytocategorizethemoviesbasedonitsmetadata, suchasreleaseyear,language,director,orcast.Mostonline video-streamingservices[3].Themainmottoofthecurrent recommendation system is formulated the basic fact that

“share”and“learn”.Theprocessofsharingandwitnessing each other’s opinions is considered as the heart of the recommendingsystem[4].Becauseofitsabilitytoprovide enhanced entertainment, a movie recommendation is becomingincreasinglypopularasapartofoursociallives.

1.1 RECOMMENDAION SYSTEM

ARecommendationsystemisatypeofinformationfiltering systemwhichisusedtopredictthe“rating”or“preference”a user would give to an item [5]. The ability to predict user preferences and needs when analyzing user behavior or otheruserbehaviortoproduceapersonalizedrecommender isaveryhelpfulfeatureinarecommendationsystem

Fig - 1: Recommendationsystem[6]

Recommendationsystemaretypicallyclassifiedintothe followingfilteringtechniqueare:

Collaborativefiltering

Content-basedfiltering

Hybridfiltering.

1.1.1 Collaborative Filtering

Themajorityofrecommendationsystemsusecollaborative filteringtoidentifysimilaruserpatternsorinformation;this method can exclude items that users like based on the ratingsorreactionsofsimilarusers.

International Research Journal of Engineering and Technology (IRJET) e-ISSN:2395-0056

Volume: 09 Issue: 11 | Nov 2022 www.irjet.net p-ISSN:2395-0072

1.1.3 Hybrid Filtering

Whichcanbethoughtofasahybridofthecontent- based andcollaborativefilteringmethods.Itavoidstheflawsofall recommendertechniques.

Fig – 2 CollaborativeFilteringtechnique[7]

Cooperativefilteringcomesintwoflavors:

A. User-based:measuresthedegreetowhichtarget usersandotherusersarecomparable.

B. Item-Based:Thismethodassesseshowsimilar twoitemsarethattargetuser’srateorinteract with.

1.1.2 Content-Based Filtering

Accordingtotheuser'spastbehaviororexplicitfeedback, content-based filtering uses item features to suggest additionalitemsthataresimilartowhattheyalreadylike.

1.2

Fig – 4 HybridFilteringtechnique[10]

Problems Related to the Recommendation System

There are some problems related to the recommendation systemareAccordingto[9,11,12].

A. Cold-start problem: when a user registers for thefirsttime,he/shenotwatchedanymovie.So, therecommendationsystemdoesnothaveany moviebasedonwhichitcangiveresult[9].This calledcold-startproblem.

B. Data scarcity problem: This problem occurs whentheuserhasratedveryfewitemsbasedon which it is difficult for the recommendation systemtoaccurateresult[9].

C. Scalability: Inthis,theencodinggoeslinearlyon items. The system works efficiently when the datasetisoflimitedsize[9].

1.3. Machine Learning Algorithms

Fig – 3 Content-basedFilteringtechnique[8]

The similarity between item vectors can be computed by threemethods[9]:

A. Cosinesimilarity

B. Euclidiandistance

C. Pearson’scorrelation

Whichfocusesonusingdataandalgorithmstomimichow humanslearn,graduallyimprovingitsaccuracy Thereare three types of machine learning techniques: supervised, unsupervised, and Semi-supervisedand reinforcement learning The machine learns under supervision in Supervised Learning. It includes a model that can predict withthehelpofalabelleddataset.Alabelleddatasetisone inwhichthetargetanswerisalreadyknown.Thefollowing supervised learning algorithms are widely used: decision tree, logistic regression, support vector machine, and knearest neighbors. Unsupervised Learning involves the machine learning on its own using unlabeled data. The machineattemptstofindapatternintheunlabeleddataand

International Research Journal of Engineering and Technology (IRJET) e-ISSN:2395-0056

Volume: 09 Issue: 11 | Nov 2022 www.irjet.net p-ISSN:2395-0072

responds.K-meansclustering,Hierarchicalclustering,and the apriority algorithm are the most commonly used unsupervisedlearningalgorithms.

A. K-Means Algorithm

Clustering[8]isaprocesstogroupasetofobjectsin such a way that objects in the same clusters are more similar to each other than to those in other clusters.

Unsupervised learning algorithm K-Means Clustering divides the unlabeled dataset into various clusters. Here, K specifies how many predefined clusters must be created as part of the process; for example, if K=2, there will be two clusters,ifK=3,therewillbethreeclusters,andso on.

classification algorithm, based on the assumption thatsimilarpointscanbefoundnearby.

Fig – 6 K-NearestNeighbors[14]

C. Support Vector Machine

ASupportVectorMachine"(SVM)canbeappliedto classification or regression problems. However, classificationissuesarewhereitismostfrequently used. When using the SVM algorithm, each data point is represented as a point in n-dimensional space,witheachfeature'svaluebeingthevalueofa specificcoordinate.Next,weperformclassification by identifying the hyper-plane that effectively distinguishes.

Fig – 5 K-meansclustering[13]

Differenttypesofdistancemeasures,includingthe Euclidean distance measure, are supported by KMeansclustering.

•TheManhattandistance

•AEuclideandistancesmeasurementofdistance

Cosinedistancecalculation

1) Euclidean distances: The most common applicationiscalculatingthedistancebetween twopoints.IfwehavetwopointsPandQ,the Euclidean distance is a straight line. It is the distance in Euclidean space between two points.FormulaforEuclideandistanceis:

B. K-Nearest Neighbors

Thek-nearest neighboralgorithm,alsoknownas KNN or k-NN, is a non-parametric, supervised learningclassifierthatusesproximitytoclassifyor predictthegroupingofasingledatapoint.Whileit can be used for either regression or classification problems, it is most commonly used as a

Fig – 7 SupportVectorMachine[15]

2. LITERATURE REVIEW

Ahmed,M.,Imtiaz,M.T.,&Khan,R.[16]discussedaboutthe recommend movie using clustering algorithm to separate similarusers&creatinganeuralnetworkforeachclustering. The result of system showed 95% accuracy on average in predictingratingfromnewuserdatawhichcanbeusedto analyzewhichmovieshouldberecommendedtonewusers.

Meshram,N.G.,etal.[17]discussedabouttheaudiofeature thatcanbeusefulintheanalysisofaffectiverepresentedin

International Research Journal of Engineering and Technology (IRJET) e-ISSN:2395-0056

Volume: 09 Issue: 11 | Nov 2022 www.irjet.net p-ISSN:2395-0072

the movie scenes and compares all methods that can be utilizedfortherecommendationofmoviesbasedonuser's emotionsexampleforSVM,SVR&HMMareanalyzedinthis paper. SVR is the best option for the selection of the connotative feature selection and mapping for the recommendationofmoviescenes.

Asad,K.I.,etal.[18]discussedaboutclassificationschemeof pre-releasemoviepopularitybasedoninherentattributes using C4.5 and PART classifier algorithm and define the relation between attributes of post release movie using correlationcoefficient.Bothofthemcorrectlyclassified77% instancesandmisclassified23%instances.

Indira,K.,andKavithadevi,M.K.[19]discussedthe(NPCAHAC)method,andamovielensdatasetwaspre-processed beforefeatureselectionusingPCA,andtheselectedfeatures wereclusteredusingk-meansandtheHACalgorithm.Based on the evaluation results, a performance analysis was performed for this ranked output, and an efficient ranked outputwithhigheraccuracyofratingandsearchresultwas obtained.Thisanalysisrevealedthattheproposedmethod outperformedtheexistingmethodologies.

Chauhan, S., [20] discussed about facial expression using convolution neural network for movie recommendation system. CNN seems more appropriate to obtain the best possibleaccuracy.Comparetorecommendationtechniques

CNN99.81%accurateresult.

NPavithaetal.[21]discussedaboutdescribesanapproach toamovierecommendationsystemusingCosineSimilarity torecommendsimilarmoviesbasedontheonechosenby the user. In this paper discussed two supervised machine learningalgorithmsNaïveBayes(NB)ClassifierandSupport Vector Machine (SVM) Classifier are used to increase the accuracyandefficiency.Thispaperalsogivesacomparison between NB and SVM on the basis of parameters like Accuracy,Precision,RecallandF1Score.Theaccuracyscore ofSVMcameouttobe98.63%whereasaccuracyscoreofNB is 97.33%. Thus, SVM outweighs NB and proves to be a betterfitforSentimentAnalysis.

3. PROPOSED SYSTEM

Our proposed system employs a number of methods, including the collaborative filtering, SVM, KNN. For the purposeofcompareSVMandKNNforbetterrecommended themovie,ourproposedmodelwillfocusoncollaborative filteringandwillbrieflyintroducetwoalgorithms thatfall intothiscategory,KNearestNeighbor(KNN)andSingular ValueDecomposition(SVD).

Step 1: Thefirststepistousethekagglemovielensdataset topredictrecommendedmoviesforeachindividual.

Step 2 The second step is carried out by employing two techniques to investigate the data at hand. To see a more comprehensiveEDAguide.

Step 3: The third step would be to introduce two collaborative-based filtering algorithms for matrix decomposition,KNNandSVM.

Step 4: Thefourthstepistoevaluatethemodelandapply the egression evaluation metrics to our recommendation system.

Fig – 8 ProposedModel

4. CONCLUSION

Inthispaper,weusecollaborativefilteringtoimplementa movie recommendation system. It is clear from the recommendationthattheresearchprojecthelpsusersexert lesseffort.Intheproposedsystem,asystemforsuggesting movies to users is built using collaborative filtering to predicthigherratings.

International Research Journal of Engineering and Technology (IRJET) e-ISSN:2395-0056

Volume: 09 Issue: 11 | Nov 2022 www.irjet.net p-ISSN:2395-0072

REFERENCES

[1] Lin, C. H., & Chi, H. (2019, March). A novel movie recommendationsystembasedoncollaborativefiltering and neural networks. In International Conference on AdvancedInformationNetworkingandApplications (pp. 895-903).Springer,Cham.

[2] Wu,C.S.M.,Garg,D.,&Bhandary,U.(2018,November). Movie recommendation system using collaborative filtering. In 2018 IEEE 9th International Conference on Software Engineering and Service Science (ICSESS) (pp. 11-15).IEEE.

[3] Kumar, S., De, K., & Roy, P. P. (2020). Movie recommendationsystemusingsentimentanalysisfrom microbloggingdata. IEEETransactionsonComputational Social Systems, 7(4),915-923.

[4] Subramaniam, R., Lee, R., & Matsuo, T. (2017, December). Movie Master: Hybrid Movie Recommendation.In 2017 International Conference on Computational Science and Computational Intelligence (CSCI) (pp.334-339).IEEE.

[5] Ahuja, R., Solanki, A., & Nayyar, A. (2019, January). MovierecommendersystemusingK-Meansclustering and K-Nearest Neighbor. In 2019 9th International Conference on Cloud Computing, Data Science & Engineering (Confluence) (pp.263-268).IEEE.

[6] S. (2022, August 12). System Design Interview: Recommendation System Design (As Used By YouTube, Netflix etc.). Medium. Retrieved November 24, 2022, from https://medium.com/double-pointer/systemdesign-interview-recommendation-system-design-asused-by-youtube-netflix-etc-c457aaec3ab

[7] Grimaldi, E. (2019, June 25). How to build a contentbasedmovierecommendersystemwithNaturalLanguage Processing. Medium.https://towardsdatascience.com/how-tobuild-from-scratch-a-content-based-movierecommender-with-natural-language-processing25ad400eb243

[8] Moshksar,R.(2021,March13). Recommender systems: context-based & hybrid recommender systems. Reza Moshksar.https://omnidatatec.com/recommendersystems-context-based-hybrid-recommender-systems/

[9] Singh, R. H., Maurya, S., Tripathi, T., Narula, T., & Srivastav, G. (2020). Movie recommendation system usingcosinesimilarityandKNN. InternationalJournalof Engineering andAdvancedTechnology, 9(5),556-559.

[10] What are the types of recommender systems? (n.d.). Quora. Retrieved November 24, 2022, from https://www.quora.com/What-are-the-types-ofrecommender-systems

[11] Mahata,A.,Saini,N.,Saharawat,S.,&Tiwari,R.(2016, December). Intelligent movie recommender system usingmachinelearning.In International Conference on Intelligent Human Computer Interaction (pp. 94-110). Springer,Cham.

[12] Ghazanfar, M., & Prugel-Bennett, A. (2010). Building switchinghybridrecommendersystemusingmachine learning classifiers and collaborative filtering. IAENG International Journal of Computer Science, 37(3).

[13] K-Means Clustering Algorithm - Javatpoint. (n.d.). www.javatpoint.com. Retrieved November 24, 2022, fromhttps://www.javatpoint.com/k-means-clusteringalgorithm-in-machine-learning

[14] K-Nearest Neighbor (KNN) Algorithm for Machine Learning - Javatpoint. (n.d.). www.javatpoint.com. Retrieved November 24, 2022, from https://www.javatpoint.com/k-nearest-neighboralgorithm-for-machine-learning

[15] Boateng,E.Y.,Otoo,J.,Abaye,D.A.,&Publishing,S.R. (2020, September 29).Basic Tenets of Classification Algorithms K-Nearest-Neighbor, Support Vector Machine,RandomForestandNeuralNetwork:AReview. Basic Tenets of Classification Algorithms K-NearestNeighbor,SupportVectorMachine,RandomForestand Neural Network: A Review. Retrieved November 24, 2022, from https://www.scirp.org/journal/paperinformation.aspx? paperid=104256

[16] Ahmed, M., Imtiaz, M. T., & Khan, R. (2018, January). Movie recommendation system using clustering and pattern recognition network.In 2018 IEEE 8th Annual Computing and Communication Workshop and Conference (CCWC) (pp.143-147).IEEE.

[17] Meshram, N. G., & Bhagat, A. P. (2014, February). Connotative features based affective movie recommendationsystem.In InternationalConferenceon Information Communication and Embedded Systems (ICICES2014) (pp.1-7).IEEE.

[18] Asad, K. I., Ahmed, T., & Rahman, M. S. (2012, May). Moviepopularityclassificationbasedoninherentmovie attributesusingC4.5,PARTandcorrelationcoefficient. In 2012 International Conference on Informatics, Electronics & Vision (ICIEV) (pp.747-752).IEEE.

International Research Journal of Engineering and Technology (IRJET) e-ISSN:2395-0056 Volume: 09 Issue: 11 | Nov 2022 www.irjet.net p-ISSN:2395-0072

[19] Indira,K.,&Kavithadevi,M.K.(2019).Efficientmachine learningmodelformovierecommendersystemsusing multi-cloud environment. Mobile Networks and Applications, 24(6),1872-1882.

[20] Chauhan, S., Mangrola, R., & Viji, D. (2021, April). AnalysisofIntelligentmovierecommendersystemfrom facialexpression.In 2021 5th International Conference on Computing Methodologies and Communication (ICCMC) (pp.1454-1461).IEEE.

[21] Pavitha,N.,Pungliya,V.,Raut,A.,Bhonsle,R.,Purohit,A., Patel, A., & Shashidhar, R. (2022). Movie Recommendation and Sentiment Analysis Using MachineLearning. Global Transitions Proceedings.