vol. 7, no. 3, 2016 a context-aware recommender context-aware recommender system for personalized...

Download Vol. 7, No. 3, 2016 A Context-Aware Recommender Context-Aware Recommender System for Personalized Places in Mobile Applications Soha A.El-MoemenMohamed Demonstrator InformationSystem

Post on 18-May-2018




0 download

Embed Size (px)


  • A Context-Aware Recommender System forPersonalized Places in Mobile Applications

    Soha A.El-Moemen Mohamed Demonstrator

    Information System Department, Faculty of Computer and Information

    Assiut University, Egypt

    Taysir Hassan A.Soliman Associate Professor

    Information System Department, Faculty of Computer and Information

    Assiut University, Egypt

    Adel A.SewisyProfessor

    Computer Science Department, Faculty of Computer and Information

    Assiut University, Egypt

    AbstractSelecting the most appropriate places under differ-ent context is important contribution nowadays for people whovisit places for the first time. The aim of the work in this paperis to make a context-aware recommender system, which recom-mends places to users based on the current weather, the time ofthe day, and the users mood. This Context-aware RecommenderSystem will determine the current weather and time of the dayin a users location. Then, it gets places that are appropriate tocontext state in the users location. Also, Recommender systemtakes the current users mood and then selects the location thatthe user should go. Places are recommended based on what otherusers have visited in the similar context conditions. Recommendersystem puts rates for each place in each context for each user.The places rates are calculated by The Genetic algorithm, basedon Gamma function. Finally,mobile application was implementedin the context-aware recommender system.

    Keywordsrecommender system, context, context-aware, ge-netic algorithm, gamma function.


    The massive scale of data on the Internet today makesit difficult for users to find relevant information. Thus, theinformation is customized according to the users needs andpreferences. Often the application of recommender systemsuses Collaborative filtering and content of the list. In ascenario of mobile devices, customization of information ismore important, because of the restrictions on mobile devicesabout the displays and input capabilities, bandwidth, etc.For example, the traveler with PDA, smart phones or pocketPCs needs access to the weather at the destination or have arecommendation for a hotel in the neighboring region. It isrecommended to customize not only the use of user profilesin advance, but also a context such as the current location,time of the day, the current weather, or the mood of theuser. Context can defined as the information that can beused to describe the situation entities i.e., person, place, orsubject, which is relevant to the interaction between the userand the application, including the user and the applicationthemselves [8]. The first challenge to generate contextrecommendations is learned via identifying the contextualfactors such as weather that affect the ratings and thus thatare worth considering. Second, get a representative set inthe context of Rates ,i.e., rankings under different contextualcircumstances is free from the context rankings. Finally, theexpansion of conventional recommender systems is to exploit

    the additional information fields in the evaluation context. Inthis paper, a novel algorithm was performed , called Context-Aware Genetic Recommender System (CAGRS). CAGRSsystem consists of five modules: 1-Collecting Stay Points,2-Clustering Stay Points, 3-Determining Popular Places ineach cluster, 4- Getting the weather and time of the day ineach Stay Point, and 5- Getting frequency of the weatherand time of the day in each cluster. CAGRS take usersparameters, such as (current users location, users mood) anddetermine the current weather and time of the day. CAGRSuses Genetic Algorithm for predicting rates of each placebased on gamma function as the fitness function. GeneticAlgorithm (GA) considered as the optimization an algorithmwhich gives an optimal solution to most problems. In thefield of Artificial Intelligence, GA is a search heuristic thatuses the process of natural selection (survival of the fittest)[9]. Genetic algorithm puts rates in each place according todifferent context conditions and for each user.The contributions of this paper are:

    1- Developing a context-aware recommender system based ona Genetic algorithm with gamma function.

    2- Rating places by using number of visits for many usersto each place and by examining context conditions in each one.

    3- Recommending the best place that user should gobased on the users mood or the current weather and time ofthe day in users location.

    This paper is divided as follows: Section 2 containssummarization of related work; Section 3 contains steps ofhow the system runs; Section 4 discusses experimental resultsand discussions. Finally, Section 5 concludes the paper andpresents future works.


    Yong Zheng and Bamshad Mobasher [19] proposed contextmatch as a new contextual modeling technique and studydifferent ways to represent context similarity and include itinto the recommendation. They showed how context similaritycan be incorporated into the sparse linear method and matrixfactorization algorithms.

    (IJACSA) International Journal of Advanced Computer Science and Applications,

    Vol. 7, No. 3, 2016

    442 | P a g ewww.ijacsa.thesai.org

  • Qiang Liu and ShuWu [6] proposed Contextual Operating Ten-sor (COT) Model, which denoted the public semantic effects ofcontexts as a contextual operating tensor and denotes a contextas a hidden vector. Then, to model the semantic operation ofa context grouping, they created contextual operating matrixfrom the contextual operating tensor and hidden vectors ofcontexts. Thus, hidden vectors of users and items can be usedby the contextual operating matrices.Wolfgang Woerndl and Christian Schueller [12] recommendedmobile applications to users based on what other users haveused in the same context. The idea is to make a hybridrecommender system to manage the extra complexity ofcontext. Users can choose among different content-based orcollaborative filtering factors, including a rule-based module,using information on point-of-interests in the vicinity of theuser, and a factor for the mixing traditional collaborativefiltering.Sabri Boutemedjet and Djemel Ziou [2] proposed a newframework for the recommendation of context-aware visualdocuments by modeling the needs of users, and context,as well as the collection of visual document together in aunified model. It also directed the user need for a variety ofrecommendations.Matthias Braunhofer, Mehdi Elahi, and Francesco Ricci [3]presented a new context-aware mobile recommender systemfor places of interest (POIs). Unlike current systems, whichstudy users preferences solely from their earlier ratings, italso considers their personality - using the Five Factor Model.Personality developed by asking users to complete a shortand entertaining questionnaire as part of the registration theprocess, and then used in: (1) an active learning elementactively acquires ratings-in-context for POIs that users arelikely to have practiced, hence decreasing the stress andannoyance to rate (or skip rating) items that the users dontknow, and (2) in the recommendation model that builds basedon matrix factorization and, therefore, can deal with unrateditems.Keith Cheverst and Nigel Davies[13] defined their skills ofdeveloping and estimating GUIDE, an intelligent electronictourist guide. The GUIDE system has built to overwhelmedmany of the limits of the old information and navigationtools available to city visitors. For example, group-based toursare inherently inflexible with fixed starting times and staticdurations and are constrained by the need to satisfy theinterests of the benefits rather than the specific interests ofindividuals. Following a period of needs capture, involvingexperts in the field of tourism, they developed and installed aSystem for use by visitors to Lancaster. The system combinesmobile computing technologies with a wireless infrastructureto present city visitors with information tailored to both theirpersonal and environmental contexts.Barry Brown and Matthew Chalmers [4] described as a lightweight portable system designed for the exchange of enter-tainment. This system allowed Visitors to the city to shareexperiences with others, through disc Computers that sharepictures, audio, and location. Collaborative filtering algorithmused historical data in previous visits that use images, the webpages and recommended places for visitors, bringing the mediatogether online with the citys streets.Pedro G. Campos and Ignacio Fernandez-Tobias [5] conductedan empirical comparison of several pre-filtering, post-filteringand contextual modeling approaches on the movie recommen-

    dation field. To acquire confident contextual information, theymade a user questionnaire where participants were asked torate movies, stating the time and social companion with whichthey preferred to watch the rated movies.Chein-Chung Hwang1, Yi-Ching Su [11] described a newrecommender system, which performed a genetic algorithm tolearn personal preferences of customers and provide tailoredsuggestions.Karthik Srinivasa Gopalan, Senthil Nathan developed the useof context capture on the users devices as a method oflearning all the user activity patterns and using these patternsto generate content recommendations. They proposed a new[9]recommendation system based on an evolutionary algorithmthat evaluates new content based on multiple objectives.Daniel Siewiorek, Asim Smailagic [1] developed SenSaywhich is a context-aware mobile phone that adjusts to dy-namically changing environmental and physiological shapes.In addition to operating ringer volume, vibration, and phonealerts, SenSay can provide faraway callers with the abilityto communicate the urgency of their calls, make call offersto users when they are idle, and provide the caller withreaction on the current status of the SenSay user. Some sensorsi


View more >