Use the GDC 2021 Session Viewer to browse and sort the ever-growing list of sessions by pass type, topic, and format. You will be able to build your personal schedule directly within the event platform, once live early July 2021.
David Renaudie (Lead Data Scientist, Massive Entertainment)
Pass Type: All Access Pass, Summits Pass
Vault Recording: TBD
Audience Level: N/A: I prefer to present live on-site but I am open to delivering a virtual presentation
Understanding the builds effectively used by players in RPG looter online games is key for designers and other stakeholders; but due to combinatorial explosion inherent to the nature of data, traditional unsupervised machine learning approaches often fail at extracting meaningful and actionable results.
We propose a novel and original method - called "BaT" - to automatically segment high volumes of player in-game data into meaningful player builds, that produce easily interpretable builds clusters in a fast, efficient, and scalable way.
Applying unsupervised machine learning to segment in-game player data into meaningful player builds and understand where is the new meta, is not an easy challenge.
Attendees learn about a novel & original approach that is able to perform this task and produce understandable & actionable results for non-ML literate audience.
Game designers, game analysts, data scientists, game developers, as well as general management.
The presentation focuses on the problematic and provides high-level highlights on the core method, as well as concrete examples from a live game-as-a-service.