Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Reinforcement learning for content's customization: a first step of experimentation in Skyscanner

Reinforcement learning for content's customization: a first step of experimentation in Skyscanner The aim of the paper is to test and demonstrate the potential benefits in applying reinforcement learning instead of traditional methods to optimize the content of a company's mobile application to best help travellers finding their ideal flights. To this end, two approaches were considered and compared via simulation: standard randomized experiments or A/B testing and multi-armed bandits.Design/methodology/approachThe simulation of the two approaches to optimize the content of its mobile application and, consequently, increase flights conversions is illustrated as applied by Skyscanner, using R software.FindingsThe first results are about the comparison between the two approaches – A/B testing and multi-armed bandits – to identify the best one to achieve better results for the company. The second one is to gain experiences and suggestion in the application of the two approaches useful for other industries/companies.Research limitations/implicationsThe case study demonstrated, via simulation, the potential benefits to apply the reinforcement learning in a company. Finally, the multi-armed bandit was implemented in the company, but the period of the available data was limited, and due to its strategic relevance, the company cannot show all the findings.Practical implicationsThe right algorithm can change according to the situation and industry but would bring great benefits to the company's ability to surface content that is more relevant to users and help improving the experience for travellers. The study shows how to manage complexity and data to achieve good results.Originality/valueThe paper describes the approach used by an European leading company operating in the travel sector in understanding how to adapt reinforcement learning to its strategic goals. It presents a real case study and the simulation of the application of A/B testing and multi-armed bandit in Skyscanner; moreover, it highlights practical suggestion useful to other companies. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Industrial Management & Data Systems Emerald Publishing

Reinforcement learning for content's customization: a first step of experimentation in Skyscanner

Download PDF
18 pages

Loading next page...
 
/lp/emerald-publishing/reinforcement-learning-for-content-s-customization-a-first-step-of-dpr5MueYJw
Publisher
Emerald Publishing
Copyright
© Emerald Publishing Limited
ISSN
0263-5577
DOI
10.1108/imds-12-2019-0722
Publisher site
See Article on Publisher Site

Abstract

The aim of the paper is to test and demonstrate the potential benefits in applying reinforcement learning instead of traditional methods to optimize the content of a company's mobile application to best help travellers finding their ideal flights. To this end, two approaches were considered and compared via simulation: standard randomized experiments or A/B testing and multi-armed bandits.Design/methodology/approachThe simulation of the two approaches to optimize the content of its mobile application and, consequently, increase flights conversions is illustrated as applied by Skyscanner, using R software.FindingsThe first results are about the comparison between the two approaches – A/B testing and multi-armed bandits – to identify the best one to achieve better results for the company. The second one is to gain experiences and suggestion in the application of the two approaches useful for other industries/companies.Research limitations/implicationsThe case study demonstrated, via simulation, the potential benefits to apply the reinforcement learning in a company. Finally, the multi-armed bandit was implemented in the company, but the period of the available data was limited, and due to its strategic relevance, the company cannot show all the findings.Practical implicationsThe right algorithm can change according to the situation and industry but would bring great benefits to the company's ability to surface content that is more relevant to users and help improving the experience for travellers. The study shows how to manage complexity and data to achieve good results.Originality/valueThe paper describes the approach used by an European leading company operating in the travel sector in understanding how to adapt reinforcement learning to its strategic goals. It presents a real case study and the simulation of the application of A/B testing and multi-armed bandit in Skyscanner; moreover, it highlights practical suggestion useful to other companies.

Journal

Industrial Management & Data Systems – Emerald Publishing

Published: Jun 7, 2021

Keywords: Reinforcement learning; Multi-armed bandit; A/B testing; Tourism and technology; Customers' need

References

  • Modern marketing: what it is, what it isn't and how do it
    Amstrong, S.; Esber, D.; Heller, J.; Timelin, B.
  • Emotionally motivated reinforcement learning based controller
    Ayesh, A.
  • Simulating sellers in online exchanges
    Bandyopadhyay, S.; Rees, J.; Barron, J.M.
  • Reinforcement learning applied to airline revenue management
    Bondoux, N.; Nguyen, A.Q.; Fiig, T.; Acuna-Agost, R.
  • Learning to plan in high dimensions via neural exploration-exploitation trees
    Chen, B.; Dai, B.; Lin, Q.; Ye, G.; Liu, H.; Song, L.
  • Marketing malpractice. Make sure all your products are profitable
    Christensen, C.M.; Cook, S.; Hall, T.
  • The use of social media in travel information search
    Chung, N.; Koo, C.
  • Adoption of travel information in user-generated content on social media: the moderating effect of social presence
    Chung, N.; Han, H.; Koo, C.
  • The influence of tourism website on tourists' behavior to determine destination selection: a case study of creative economy in Korea
    Chung, N.; Lee, H.; Lee, S.J.; Koo, C.
  • Open sources deep mind lab
  • Predicting rank for scientific research papers using supervised learning
    El Mohadab, M.; Bouikhalene, B.; Safi, S.
  • Model-based learning protects against forming habits
    Gillan, C.M.; Otto, A.R.; Phelps, E.A.; Daw, N.D.
  • When user experience designers partner with data scientists
    Girardin, F.; Lathia, N.
  • Reinforcement learning explained
    Hu, J.
  • Second-best beam-alignment via bayesian multi-armed bandits
    Hussain, M.; Michelusi, N.
  • Online experimentation at Microsoft
    Kohavi, R.; Crook, T.; Longbotham, R.; Frasca, B.; Henne, R.; Ferres, J.L.; Melamed, T.
  • Algorithms for the multi-armed bandit problem
    Kuleshov, V.; Precup, D.
  • Semi-Markov adaptive critic heuristics with application to airline revenue management
    Kulkarni, K.; Gosavi, A.; Murray, S.; Grantham, K.
  • A short introduction to learning to rank
    Li, H.
  • A contextual-bandit approach to personalized news article recommendation
    Li, L.; Chu, W.; Langford, J.; Schapire, R.E.
  • Learning to rank for information retrieval
    Liu, T.Y.
  • Explore, exploit, and explain: personalizing explainable recommendations with bandits
    McInerney, J.; Lacker, B.; Hansen, S.; Higley, K.; Bouchard, H.; Gruson, A.; Mehrotra, R.
  • Artwork personalization at Netflix
    Medium
  • Can machine learning techniques predict customer dissatisfaction? A feasibility study for the automotive industry
    Meinzer, S.; Jensen, U.; Thamm, A.; Hornegger, J.; Eskofier, B.M.
  • Dynamic online pricing with incomplete information using multiarmed bandit experiments
    Misra, K.; Schwartz, E.M.; Abernethy, J.
  • Reinforcement learning on a futures market simulator
    Moriyama, K.; Matsumoto, M.; Fukui, K.I.; Kurihara, S.; Numao, M.
  • Cas-GANs: an approach of dialogue policy learning based on GAN and RL techniques
    Nabeel, M.; Riaz, A.; Zhenyu, W.
  • A hybrid method with TOPSIS and machine learning techniques for sustainable development of green hotels considering online reviews
    Nilashi, M.; Mardani, A.; Liao, H.; Ahmadi, H.; Manaf, A.A.; Almukadi, W.
  • A machine learning-based usability evaluation method for eLearning systems
    Oztekin, A.; Delen, D.; Turkyilmaz, A.; Zaim, S.
  • Learning diverse rankings with multi-armed bandits
    Radlinski, F.; Kleinberg, R.; Joachims, T.
  • Reinforcement learning agents
    Ribeiro, C.H.C.R.
  • Some aspects of the sequential design of experiments
    Robbins, H.
  • Learning and decisions in contextual multi-armed bandit tasks
    Schulz, E.; Konstantinidis, E.; Speekenbrink, M.
  • Mastering the game of go without human knowledge
    Silver, D.; Schrittwieser, J.; Simonyan, K.; Antonoglou, I.; Huang, A.; Guez, A.; Hubert, T.; Baker, L.; Lai, M.; Bolton, A.; Chen, Y.; Lillicrap, T.; Hui, F.; Sifre, L.; van den Driessche, G.; Graepel, T.; Hassabis, D.
  • Model-based reinforcement learning with nearly tight exploration complexity bounds
    Szita, I.; Szepesvári, C.
  • Multi–armed bandit for pricing
    Trovo, F.; Paladino, S.; Restelli, M.; Gatti, N.
  • Turn customer input into innovation
    Ulwick, A.W.
  • Contextual: evaluating contextual multi-armed bandit problems in R
    van Emden, R.; Kaptein, M.
  • Airdialogue: an environment for goal-oriented dialogue research
    Wei, W.; Le, Q.; Dai, A.; Li, J.
  • Enhancing collaborative filtering music recommendation by balancing exploration and exploitation
    Xing, Z.; Wang, X.; Wang, Y.
  • From infrastructure to culture: A/B testing challenges in large scale social networks
    Xu, Y.; Chen, N.; Fernandez, A.; Sinno, O.; Bhasin, A.
  • Credit risk assessment using machine learning techniques
    Aithal, V.; Jathanna, R.D.
  • Customer relationship management and big data enabled: personalization and customization of services
    Anshari, M.; Almunawar, M.N.; Lim, S.A.; Al-Mudimigh, A.
  • Semi-supervised learning based aesthetic classifier for short animations embedded in Web pages
    Bansal, D.; Bhattacharya, S.
  • Supporting customer-oriented marketing with artificial intelligence: automatically quantifying customer needs from social media
    Kühl, N.; Mühlthaler, M.; Goutier, M.
  • Airline itinerary choice modeling using machine learning
    Lhéritier, A.; Bocamazo, M.; Delahaye, T.; Acuna-Agost, R.
  • A new machine learning method for identifying Alzheimer's disease
    Liu, L.; Zhao, S.; Chen, H.; Wang, A.
  • Predicting the helpfulness of online restaurant reviews using different machine learning algorithms: a case study of yelp
    Luo, Y.; Xu, X.
  • Is my interface beautiful?—a computational model-based approach
    Maity, R.; Bhattacharya, S.
  • Machine learning analysis of mortgage credit risk
    Pillai, S.G.; Woodbury, J.; Dikshit, N.; Leider, A.; Tappert, C.C.
  • Managing regional consumer market based on the improved approach to evaluating customer needs for food commodities
    Soboleva, Y.P.; Parshutina, I.G.; Voronkova, I.E.; Shaporova, O.A.
  • Investigating antecedents of behavioral intentions in mobile commerce
    Song, J.; Koo, C.; Kim, Y.
  • Identifying customer needs from user-generated content
    Timoshenko, A.; Hauser, J.R.
  • Machine learning-based multiparametric MRI radiomics for predicting the aggressiveness of papillary thyroid carcinoma
    Wang, H.; Song, B.; Ye, N.; Ren, J.; Sun, X.; Dai, Z.; Zhang, Y.; Chen, B.T.
  • An approach to discovering product/service improvement priorities: using dynamic importance-performance analysis
    Wu, J.; Wang, Y.; Zhang, R.; Cai, J.