Data-Driven Destination Recommender Systems

Linus W. DietzTechnical University of MunichDepartment of InformaticsChair of Connected Mobility (I11)UMAP’18, SingaporeJuly 10, 2018

Introduction

ProblemRecommend composite trips of global travel destinations

“I want to travel to South-East Asia for sixweeks in summer to experience culture,good food and go hiking in the mountains.I have a budget of $1500.”

MotivationIndependent travel planning is complex,information is scattered, outdated, andof uncertain quality

Challenges� Find distinct touristic regions� Classification of tourist destinations� User modeling with little user effort� Recommendation algorithm

c Created by Freepik

Linus W. Dietz (TUM) 2

Overview

Items User Algorithms

� Item discovery

� Item classification

� Effective and effortlesspreference elicitation

� Traveler clustering

� Content-based filtering� Constraint satisfaction� Diversity of activities� Durations of item

consumption

Linus W. Dietz (TUM) 3

Data Mining & Domain Modeling

Recommendation items: set of travel regionsCombination of heterogeneous data sources

Item discovery

� Mismatch between political regions and tourist destinations!� Employ hierarchical region tree� Make the granularity of destinations dependent on the query area

Item characterization� What are the characteristics, attractions, and activities of a destination?� Aggregation based on single points of interests� What is the typical duration of stay at a destination?� Analyze tourist mobility patterns for domain understanding

Evaluation: Offline comparison with crowdsourcing and expert knowledge

Contribution: framework for data-driven recommender systemsLinus W. Dietz (TUM) 4

User ModelingPreference elicitation

� Which activities are best suited for a traveler?� How can traveling preferences be elicited effectively with little effort?

Traveler clustering

� What are the relevant features to characterize travelers?� What distinct types of travelers are there?� How can the pace of the travel itinerary be personalized based on past trips?

Evaluation: Controlled lab experiments

Contribution: Novel approaches for domain-specific user modeling

Linus W. Dietz (TUM) 5

Recommendation Algorithms

Content-based recommendation under constraints. Knapsack problem!

Personalize item consumption durations

Ensure sufficient diversity within a trip

Measure the benefits of explaining recommendations and critiquing

Evaluation: Measure user satisfaction in an online field study

Contribution: Constraint-based algorithms for composite trips

Linus W. Dietz (TUM) 6

Current Progress

Investigated traveler mobility patterns

Linus W. Dietz (TUM) 7

Current Progress

Preliminary investigation of durations of stay

Day

s

Kuw

ait

Dom

inic

an R

ep.

Cyp

rus

Aze

rbai

jan

Cos

ta R

ica

Gha

naK

azak

hsta

nP

arag

uay

Aus

tral

iaB

elar

usS

audi

Ara

bia

Turk

eyM

artin

ique

Bra

zil

Uni

ted

Sta

tes

Rus

sia

Col

ombi

aM

exic

oP

hilip

pine

sC

anad

aN

ew Z

eala

ndIn

done

sia

Japa

nA

rgen

tina

Chi

naU

nite

d K

ingd

omM

alay

sia

Ven

ezue

laT

haila

ndS

pain

Sin

gapo

reU

n. A

rab

Em

irate

sIta

lyIr

elan

dG

erm

any

Sw

eden

Fra

nce

Cze

ch R

ep.

Fin

land

Net

herla

nds

Sw

itzer

land

Bel

gium

Den

mar

kH

unga

ryE

ston

ia

03

69

1215

1821

050

015

0025

0035

00O

bser

vatio

ns

mean days

observations

Linus W. Dietz (TUM) 8