Tour Recommendation on Location-based Social Networks [Extended Abstract] Kwan Hui Lim

Jeffrey Chan

The University of Melbourne and NICTA

RMIT University, Australia

[email protected] Christopher Leckie

[email protected] Shanika Karunasekera

The University of Melbourne and NICTA

The University of Melbourne

[email protected]

[email protected] 1

ABSTRACT

2

20 min

Landmark (10 min)

We propose two tour recommendation algorithms based on the Orienteering problem, and implement a tour recommendation framework that uses geo-tagged photos and Wikipedia to determine the travel histories and interests of tourists.

50 min

Leisure (4 hrs) 40 min

Food (1 hr)

Travel Sequences (Visit duration, travel time, user interests)

Landmark (10 min)

Geo-tagged photos (lat/long, time taken)

3

Categories and Subject Descriptors H.2.8 [Database Management]: Database Applications Data mining; H.3.3 [Information Storage and Retrieval]: Information Search and Retrieval

List of Places of Interests (lat/long, category)

UserID

POI Popularity

Interests

684115665

{Park, Muesuem}

165514864

{Shopping, Beach}

165518648

{Food, Landmark}

123248566

{Shopping}

122365448

{Park, Beach, Food}

211568645

{Landmark}

231231654

{Muesuem, Park}

231231848

{Food, Church, Park}

323123165

{Shopping, Food}

545645645

{Beach, Park}

Tourist Interests

General Terms Figure 1: Tour Recommendation Framework

Algorithms, Experimentation, Measurement

Keywords Tour Recommendation, Travel Itinerary, User Interests, Orienteering Problem, Flickr, Wikipedia, Social Networks

1.

INTRODUCTION

For a visitor in a foreign city, it is a challenging task to plan a tour or itinerary that is customized to his/her interest preferences. Despite the availability of online resources, there are challenges in planning a customized tour, namely: (i) many travel guides recommend popular Points of Interest (POI) that are not personalized to the tourist’s interest preferences; (ii) even with a list of POIs, it is a tedious task to construct an itinerary based on those POIs with the considerations of travelling time and specific starting/ending points. Our work addresses these challenges by proposing two algorithms based on variants of the Orienteering problem and implementing a tour recommendation framework (Fig. 1) that utilizes geo-tagged photos and Wikipedia.

2.

PROPOSED APPROACH

We model the tour recommendation problem as a variant of the Orienteering problem [3], with a time/distance budget B, starting POI p1 and destination POI pN . Our goal is to recommend a tour T = (p1 , ..., pN ) that maximizes POI popularity and tourist interest, while staying within the budget B. Specifically, we proposed the following algorithms: • PersTour algorithm [2]: For recommending personalized tours with POIs and visit duration based on POI popularity and time-based user interests. • TourRecInt algorithm [1]: For recommending customized tours with a mandatory POI category based on user/tourist interests.

Our overall tour recommendation framework (Fig. 1) comprises the following steps: 1. Map geo-tagged photos (e.g. Flickr) to a list of POIs (e.g. Wikipedia) if their coordinates differ by <100m. 2. Construct the tourist travel history by connecting POI visits (obtained from Step 1) of the same tourist. 3. Calculate POI popularity and tourist interest preference based on tourist travel histories from Step 2. Thereafter, we can apply our PersTour or TourRecInt algorithms to recommend customized tours.

3.

EXPERIMENTS AND RESULTS

We evaluate the effectiveness of PersTour and TourRecInt against various greedy-based baselines on a Flickr dataset of multiple cities. Our experimental results show that PersTour and TourRecInt out-perform the baselines in terms of various evaluation metrics, namely: tour popularity, tour interest, POI count, precision, recall, F1 score, and RMSE of visit duration. For more information on our proposed algorithms and experimental results, please refer to [1] and [2]. Acknowledgments. This work was supported by NICTA.

4.

REFERENCES

[1] K. H. Lim. Recommending tours and places-of-interest based on user interests from geo-tagged photos. In Proceedings of the 2015 SIGMOD PhD Symposium (SIGMOD’15), 2015. [2] K. H. Lim, J. Chan, C. Leckie, and S. Karunasekera. Personalized tour recommendation based on user interests and points of interest visit durations. In Proc. of the 24th Intl. Joint Conf. on Artificial Intelligence (IJCAI’15), 2015. [3] T. Tsiligirides. Heuristic methods applied to Orienteering. Journal of the Operational Research Society, 35(9), 1984.

Tour Recommendation on Location-based Social Networks

INTRODUCTION. For a visitor in a foreign city, it is a challenging ... Intl. Joint Conf. on Artificial Intelligence (IJCAI'15), 2015. [3] T. Tsiligirides. Heuristic methods ...

551KB Sizes 4 Downloads 322 Views

Recommend Documents

Recommendation on Item Graphs
Beijing 100085, China [email protected] Tao Li. School of Computer Science. Florida International University. Miami, FL 33199 [email protected]

Online Video Recommendation Based on ... - Semantic Scholar
Department of Computer Science and Technology, Tsinghua University, Beijing 100084, P. R. ... precedented level, video recommendation has become a very.

Recommendation Based on Personal Preference
Oct 2, 2003 - tion at run-time according to user request and available data, then .... that is, whether a data item satisfies a condition is a matter of true or false,.