PREDICTING INDIVIDUAL PLAYER BEHAVIOR OPERATIONALLY WITH BIG DATA AND MACHINE LEARNING

Transcript

1. PREDICTING INDIVIDUAL PLAYER BEHAVIOR OPERATIONALLY WITH BIG DATA AND MACHINE LEARNING África Periáñez, PhD CEO Yokozuna Data, a Keywords Studio 4 December 2018, Tokyo A KEYWORDS STUDIO YOKOZUNAdata
2. PhD Mathematics (Ensemble Learning) - 2015 University of Reading MSc String Theory - 2006 CERN MSc Theoretical Physics - 2003 UAM BSc Physics - 2001 UAM Dr. África Periáñez CEO of Yokozuna Data
3. 2005/2006 2006/2010 2011/2015 2015/...
4. Ana Fernández, MSc SENIOR RESEARCH DATA SCIENTIST Pei Pei Chen, MSc MACHINE LEARNING ENGINEER LEAD Anna Guitart, MSc DATA SCIENTIST Cristian Conteduca, MSc BACKEND ENGINEER LEAD Vitor Santos, MA UX DESIGN & CREATIVE LEAD Nitin Kumar, MSc FULL-STACK ENGINEER Pooja Revanna, MSc BACKEND ENGINEER Peng Xiao, BSc BIG DATA ENGINEER Omid Aladini, MSc DATA INFRASTRUCTURE ADVISOR Dexian Tang, MSc BIG DATA ENGINEER Javier Grande, PhD SCIENTIFIC EDITOR Alvaro de Benito, MA PR & COMMUNICATION LEAD Shi Hui Tan, MSc DATA SCIENTIST Yokozuna Data’s team
5. + + DevOps AI CLOUD DEVOPS
6. Yokozuna Data AWS GameSparks AI Cloud DevOps
7. Founded in 2015, joined Keywords Studios in 2018 to push back the frontiers of General Behavioral Machine Learning and to revamp the video-game industry: Personalized games What is Yokozuna Data?
8. Mission To unlock the knowledge of big game databases To convert unstructured data into actionable information to understand and predict individual player behavior
9. A state-of-the-art machine learning engine that predicts individual player behaviour Pushing the game data science frontiers to a new limit:
10. ACADEMIA SCIENTIFIC RESEARCH INTERDISCIPLINARITY START-UP VIDEO GAME BUSINESS CHALLENGES
11. Highly sophisticated games allow players to express nuanced emotions through their in-game actions
12. When will players leave the game? Time Level Playtime Money
13. Which item will they purchase next? Billing history Distribution of item probabilities Time to next purchase 08/06 08/15 08/27 09/06 GAME APPLICATIONS Players who may stop purchasing Playtime Per player Upcoming churners
14. Personalization Personalized matching Who should you compete against in Mario Kart? Which clan is your best opponent in Clash of Clans?
15. Item recommendation system Action recommendation Rewards and discounts Actionable recommendations to increase engagement and retention Personalization
16. Predicting player behavioral outcomes is key to the success of game developers Video game challenges
17. Man is a deterministic device thrown into a probalistic Universe Amos Tversky Daniel Kahneman Nobel prize 2002
18. NEW PLAYERS Acquiring new users is expensive: in Japan, the average cost-per-install for gaming apps reached 6.07 USD in 20181 Video game challenges: Increasing Retention Between 75–90% of new players churn on the first day 1 Liftoff and Adjust, 2018 and inefficient: only 5% remain after one month
19. Retention of the most valuable players is crucial: the top 10% of paying users contribute 60% of the revenue Thousands of titles are published every year and compete for same playersʼ time and attention Video game challenges: Increasing Retention VIP PLAYERS
20. Only a small fraction of users make purchases, dentifying these users and predicting their Customer Lifetime Value is crucial A key challenge for game developers is to convert players from non-premium to premium Video game challenges: Maximizing the engagement of VIP players
21. Targeting the right players by customizing game events and publishing them at the right time Identify the reasons behind behavioral trends Find the best acquisition, marketing and game event strategies Video game challenges: Optimizing Game Events and Marketing Campaigns
22. Solving Video Game Business Challenges
23. Numbers for AAA mobile games $1.3M/month sales increase +10% VIP Retention $4M/month sales increase +5% PU Increase $8M/year sales increase Improving development of levels with the highest churn rate
24. Numbers for Asian medium-size RPGs $8.7K/month sales increase $22K/month sales increase +5% PU Increase $330K/year sales increase +10% VIP Retention Improving development of levels with the highest churn rate
25. 1/ Operational player behavioral prediction 2/ Deep learning and ensemble learning 3/ Big data
26. Multiple learners are trained to solve the same problem Ordinary machine learning approaches learn one hypothesis from training data Ensemble methods try to construct a set of hypotheses and combine them OUTPUT Ensemble Learning
27. label Reconstruction Error Output nth feature layer 2nd feature layer 1st feature layer Input label Learning Learning & Generalization Deep Learning Model Artificial Neural Network Deep Learning
28. CHURN PREDICTION IN MOBILE SOCIAL GAMES: TOWARDS A COMPLETE ASSESSMENT USING SURVIVAL ENSEMBLES A. Perianez, A. Saas, A. Guitart and C. Magne IEEE DSAA 2016 Montreal DISCOVERING PLAYING PATTERNS: TIME SERIES CLUSTERING OF FREE-TO-PLAY GAME DATA A. Guitart, A. Perianez and A. Saas, IEEE CIG 2016 Santorini GAMES AND BIG DATA: A SCALABLE MULTI-DIMENSIONAL CHURN PREDICTION MODEL P. Bertens, A. Guitart and A. Perianez, A. Saas, IEEE CIG 2017 New York FORECASTING PLAYER BEHAVIORAL DATA AND SIMULATING IN-GAME EVENTS A. Guitart, P. Chen, P. Bertens and A. Perianez IEEE FICC 2018 Singapore THE WINNING SOLUTION TO THE IEEE CIG 2017 GAME DATA MINING COMPETITION A. Guitart, P. Chen, A. Periáñez A MACHINE-LEARNING ITEM RECOMMENDATION SYSTEM FOR VIDEO GAMES P. Chen, A. Guitart, P. Bertens, A. Periáñez IEEE CIG 2018 Maastricht Yokozuna Data peer-reviewed articles CUSTOMER LIFETIME VALUE IN VIDEO GAMES USING DEEP LEARNING AND PARAMEDIC MODELS P. Chen, A. Guitart, A. Fernandez de Rio and A. Periáñez IEEE Big Data 2018 Seattle yokozunadata.com/research/ A KEYWORDS STUDIO YOKOZUNAdata
29. Milestones International press coverage by top MEDIA Best Paper Award at FICC2018 1 Winners of game churn prediction CONTEST NCSoft CMU aplicants for summer internship +40 7 peer-reviewed articles published in top machine learning conferences
30. THE WINNING SOLUTION TO THE IEEE CIG 2017 GAME DATA MINING COMPETITION
31. Number of registrants in the competition 2 tracks: Which players will leave the game 264 Teams When they will leave the game
32. Results Rank 1 2 3 4 5 YokozunaData (Japan) UTU (Finland) TripleS (Korea) TheCowKing goedleio 0.610098 0.60326 0.57968 0.59370 0.57717 0.63326 0.60370 0.62459 0.60718 0.56205 0.62145 0.60348 0.60130 0.60036 0.58882 Team Test1 Score Test2 Score Total Score Rank 1 2 3 4 5 YokozunaData (Japan) IISLABSKKU UTU (Finland) TripleS (Korea) DTND 0.883248 1.034321 0.927712 0.958308 1.032688 0.616499 0.679214 0.898471 0.891106 0.930417 0.726151 0.819972 0.912857 0.923486 0.978888 Team Test1 Score Test2 Score Total Score Track 1 Which players will leave the game Track 2 When they will leave the game
33. An Operational Prediction System Big Data Engineering Infrastructure A KEYWORDS STUDIO YOKOZUNAdata
34. The Problem SCALING TO INFINITY The Solution Supporting data from thousands of games and millions of Monthly Active Users (MAU) A CLOUD DISTRIBUTED-SYSTEM DESIGN FOR: - Data upload - Databases and storage - Parallel computing for data processing and machine learning execution
35. UPLOAD COMPUTE RESULTS Results Data Upload Servers/ Storage Compute Instances Cloud Storage Computing Servers Web Servers Database Servers Kubernetes
36. Summary The biggest challenge is to be able to retain users with very diverse tastes, desires and motivations - Profile players at the individual level and direct them towards the activities that are more likely to interest them - Tailor game events, pricing and promotions - Identify VIP players sooner to provide a premium service 1
37. Yokozuna Data provide accurate predictions of: - Individual-user actions - Personalized promotions and rewards - Moment and level at which they will leave - Money they will spend in the game and their potential Summary 2
38. Contact: aperianez@yokozunadata.com linkedin.com/in/africaperianez www.yokozunadata.com @yokozunadata A KEYWORDS STUDIO YOKOZUNAdata
39. THANK YOU! ;) A KEYWORDS STUDIO YOKOZUNAdata