Personalized Location Recommendation System Personalized Location Recommendation System: A Review

Personalized Location Recommendation System Personalized Location Recommendation System: A Review

Ashwini Arun Ughade (Dr. Babasaheb Ambedkar Marathwada University, Aurangabad, India)
Copyright: © 2019 |Pages: 10
DOI: 10.4018/IJAEC.2019010104


Location acquisition and wireless communication technologies are growing in location-based social networks. With the rapid development of location-based social networks (LBSNs), location recommendation has become an important for helping users to discover interesting locations. Most current studies on spatial item recommendations do not consider the sequential influence of locations. The authors proposed a personalized location recommendation system as a probabilistic generative model that aims to mimic the process of human decision-making when visiting locations. In this system, three tasks are involved, such as: extracting user's personal interests; extracting sequential influence; and combining them into unified networks. This system utilizes data collected from LBSNs to model a user's behavior and locations with real datasets, and it determines a user's preferred locations using collaborative filtering and a Locality Sensitive Hashing (ALSH) technique. It overcomes the challenges of the user's check-in data in LBSNs having a low sampling rate in both space and time and a huge prediction space.
Article Preview

1. Introduction

Recommendation systems have been widely used to provide users with high-quality personalized recommendations from a large check-ins data in LBSNs. Collaborative filtering (CF) is one of the most popular techniques to implement a recommender system. Location-based social networks (LBSNs) provide interface to the users for attracting their interesting locations. In recent year, location base services are growth in location-based social networks. A personalized location recommendation services encourages to the users for exploring new locations. Therefore, developing personalized recommendation systems for LBSNs to provide users with spatial items has recently attracted increased research attention where People are check our location and share daily activity in web. This system is a probabilistic generative model inspired by SAGE model that aims to mimic the process of human decision making when visiting spatial items using collaborative filtering. Personalized Location Recommendation Systems use the check-ins history of LBSNs data to find personal interest locations and use their public preferences. In sequential recommendation locations are stored in sequential manner with different time. It combines the sequential location with personal interest location in the networks. Personalized Location Recommendation System applied with collaborative filtering and Locality Sensitive Hashing (ALSH) technique for increase the location accuracy.

Personalized Location Recommendation System may analyze on real-world datasets like Foursquare, Gowalla and Brightkite these datasets are publicly available. This system works on foursquare dataset. This dataset contains the check-in history of users who live in New York, USA, between April 2012 and Nov 2012 having 227,428 records. Each check-in activity contains the user-ID, venue- ID, venue-category, venue-location and a check-in time. In location recommendation check-ins data used for calculating the non-uniform distribution of transition probabilities of each of next location in sequence visited by the users. These sequential patterns result from different factors for personal interest and public preferences of location.

Personalized spatial item recommendation explores the interesting location to the user’s for improve their recommendation accuracy, based on personal interest location and public preferences of location. In social networks, users update daily affairs with location. Users can check geographical spaces or any activity.

Sequential information for spatial item recommendation is highly challenging, mainly due to the following problems.

1.1. Low-Sampling Rate

GPS and LBSN provide the location related data. A GPS trajectory provides the large amount user’s location histories and contains the time information. Check-in history of LBSNs data collected from location based social networks like Gowalla, Brightkite having low sampling rate compared to GPS trajectories data in both space and time. It is difficult to model the dependency between two check-in LBSNs using the location prediction techniques on GPS routes.

1.2. Huge Prediction Space

In LBSN (Bao et al., 2012), check-in history used for recommendation. This system based on check in history finding the personal interest location and public preferences of location. These preferences are stored in sequence format with the time information. If user wants to personal interest locations, then collaborative filtering is used for finding other personal interest locations and public preferences locations through top-k recommendation.

In location recommendation, locations are stored in Markov chains. Finding the probability of each location in Markov chain is very difficult, time consuming and also computationally expensive process. They use sequential pattern of locations based on the first-order Markov chain for efficiency, but this only considers the newest visited place of a user. A user visiting a new location through additive Markov chain reflects the effect of all visited places in the check-in history of the user on the new location. Sequential recommendation methods have been proposed in the literature (Yin et al., 2016), most of which are based on Markov chains. To reduce the prediction space, most related studies (Zhang & Chow, 2015; Bao et al., 2012) first-order Markov chain used sequential location, which considers only the last one in a sequence of locations visited by a user to recommend a new location. Therefore, we need to develop a new method to incorporate the influence from all recently visited locations, rather than just the last one, to make location recommendations within a small predication space.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 11: 4 Issues (2020): Forthcoming, Available for Pre-Order
Volume 10: 4 Issues (2019): 3 Released, 1 Forthcoming
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing