Type: Classification Essays
Sample donated: Pat Lloyd
Last updated: April 12, 2019
Abstract: Web personalization aims to tailor the Web to provide the users what they want and how they want it, rather than providing the same content, in the same style or format, to all its diverse range of users. Interest in Web personalization is rapidly increasing and the number of commercial organizations touting to provide software and services for personalization is mushrooming.
Personalization is seen as a major advancement in the evolution of Web and Web-based applications, and is being deployed in e-business and other Web applications. Restaurant recommendation system is a very popular service whose accuracy and sophistication keeps increasing every day. With the advent of smart phones, web 2.0 and internet services like 3G, this has become accessible by every consumer.
Here, we present a personalized location based restaurant recommendation system. This recommender system adopts a user preference model by using the features of user’s visited restaurants, and also utilizes the location information of user and restaurants to dynamically generate the recommendation results. Keywords: Web personalization, recommender systems, collaborative filtering, content filtering, hybrid filtering, Naive Bayes algorithm. I.INTRODUCTION In this age of information overload, people use various methods to choose from what to buy, what to read or what to eat.
Recommendation systems automate some of these methods with the goal of providing efficient, personal and high quality recommendations Recommender systems have changed the way many sites on the internet interact with users. Instead of giving a experience which is static in nature where a user can just search for products, these systems intend to improve communication to give a better experience. Recommender systems identify recommendations autonomously for individual users based on past purchases and searches, and on another users’ behaviour. Recommendation systems are a type of information filtering that presents lists of items which are likely of user interest. Amazon, Netflix, Pandora are the most popular recommender systems all over the world. Simply they compare user interest acquired from his/her profile with some reference characteristics and predict the rating that the user would give. In this paper we have proposed the study of different features of various techniques used for making predictions in recommender systems. 1.
1 OBJECTIVE Recommendation systems have been used since a long time to filter the user’s choices. The two basic approaches that are used for recommender systems are content-based filtering and collaborative filtering. Content based filtering works by examining the characteristics of the items to display more items. Collaborative Filtering studies the past behavior and preferences of similar users. Collaborative Filtering focuses on studying the relationship between user and items.In this system, we have used the second method that is collaborative filtering and study the relationship of users and restaurants which is stored in the data by analyzing reviews,ratings with the help of Foursqaure API to build our system.
The reviews of users which are basically in the form of statistical data are available on the web.Websites like Zomato, Foursqaure and Yelp include user reviews which cover huge numbers of eateries across the globe. The reviews include information about the opinions of various users’ and their choices, which are an asset to any recommender system. AI. RELATED WORK In any e-commerce application, the recommender systems play a vital role as they assist the prospective buyers in making proper decisions on the basis of the recommendations that the system provides. Recommender systems aim at providing the users with effective recommendations based on their intuitions and preferences. The two techniquescommonly used for providing automated recommendations are collaborative filtering and knowledge based filtering techniques.
Anant Gupta and Kuldeep Singh present a personalized location based restaurant recommendation system integrated in mobile technology. It ubiquitously studies the user’s behavioural pattern of visiting restaurant using a Machine Learning algorithm. It also addresses the issues faced by today’s recommendation systems and propose methods to rectify it.
Further it can be expanded to search and update restaurants based on the user’s friend’s recommendations.This study proposes a model that combines localization, personalization and content-based recommendation in a dynamic and ubiquitous environment. A different and unsocial form of personalization that is only derived from the user’s behaviour and caters to his needs is designed. The common problems have also been addressed and to dealt with. The architecture is kept minimal and light and common algorithms are fused in a unique way. Recommender systems continue to evolve on a daily basis, and we believe we have contributed to this evolution 3.PS Lokhande and BB Mesharam proposed the class diagram for Pre-processing Systems, Knowledge Discovery Systems and Recommendation Subsystems and then provided workflow of the systems by proposing the required functional model of the systems.
In the design of the systems it proposed the data structure design and Dynamic Web Systems architecture and implemented the systems. Lastly, they have shown the results of the navigational behaviour of the user during particular transaction. The main element of a Web personalization system is the usage miner. Log analysis and Web usage mining is the procedure where the information stored in the Web server logs is processed by applying Knowledge Discovery and data mining techniques, such as clustering, association rules discovery, classification and sequential pattern discovery, in order to reveal useful patterns that canbe further analysed. 2 BI.
PROPOSED METHOD Most of the recommendation systems, especially, Knowledge Based Recommendation Systems, prompt the user constantly for category and other attributes of the restaurant. Also, the user has to rate and review items, and share his location on social networks. The user has to manually start the system whenever he needs the service. This application gives the users a top recommended restaurant based on their interests within the given time limit. The system would use push notification service to constantly recommend places to user, when he is on move. The system would automatically share his location on social networks, depending on the settings he has provided Fig 3.1 Proposed work 3.2 System Flow The system uses collaborative approach using Naive Bayes algorithm to recommend the users top restaurants.
Based on the type of functionalities, the system can be divided into different modules, i.e., database layer, recommendation engine, user profile generator and online interaction layer. Based on the user’s check-in to any restaurant, the data will be added to the restaurant data set and will then be evaluated for user interest profile. Fig 3.2 System flow A.
User LoginAfter the user has filled the personal details including the name phone number, email id, state, birth of date and location, the user logs in with his respective username and password. B. Geo-Locate UserAfter logging in, the user is asked to enter three locations nearby his current locations. The Google map geocoding API will automatically detect the current user locations and the respective locations near him/her. C. Learn user profile After locating the user, the user filters his/her type of cuisines according to the taste, the recommendation engine keeps track of the user choices and starts filtering the requests. D.
Display Restaurant near userAfter filtering the recommendation engine displays the list of restaurants according to user’s current location E. Display recommendationAlong with restaurants near user, the user also gets recommendation of the restaurants from the group of users which are according to them top rated/ favourite restaurants. F. Rate and reviewIf the user has been to the restaurant or wants to check the review of the restaurant he/she can rate or review that particular venue. G. Restaurant information updatedAfter getting the required details from users i.
e. rate or reviews from the user that particular restaurant information is updated. 3.3 Naive Bayes Classifier Algorithm It is one of the probabilistic classification method which is based on Bayes’ classifier. In other words, the algorithm assumes the presence of a specific characteristic in a class is not related to the presence of any other characteristic. For instance, a vegetable can be regarded to be tomato if it is red and round.
Naive Bayes model is easy to build and particularlyuseful for very large data sets. Along with simplicity, Naive Bayes is known to outperform even highly sophisticated classification methods3. IV. CONCLUSION This study proposes a model that combines localization, personalization and content-based recommendation in a dynamic and ubiquitous environment. A different and unsocial form of personalization that is only derived from the user’s behavior and caters to his needs is designed. The common problems have also been addressed and to dealt with.
The architecture is kept minimal and light and common algorithms are fused in a unique way. Recommender systems continue to evolve on a daily basis, and we believe we have contributed to this evolution. V. REFERENCES 1 L Flory, KM Osei-Bryson, M Thomas – Decision Support Systems, 2017 – Elsevier, “A new web personalization decision-support artifact for utility-sensitive customer review analysis”.2 PS Lokhande, B.
B Meshram,”Analysis and design of web personalization system for E- Commerce”.3 Anant Gupta,Kuldeep Singh, “Location Based Personalized Restaurant Recommendation System” 4 Anand S, Mobasher B. (2012), “Intelligent Techniques of Web Personalisation”, Department of Computer Science, University of Warwick, Coventry CV4 7AL, UK. 5A Hybrid Restaurant Recommender Prerna Dwivedi Mumbai,India Nikita Chheda 6 FoodR Recommender System for restaurants Author: Dan Claudiu MANOLI http://citeseerx.ist.psu.
edu/viewdoc/download?doi=10.1.1.244. 8702=rep1=pdf7http://www.cs.carleton.edu/cs_comps/0607/recommend/reco mmender/itembased.htm