A process for designing algorithm-based personalized gamification

Abstract

Personalization is an upcoming trend in gamification research, with several researchers proposing that gamified systems should take personal characteristics into account. However, creating good gamified designs is effort intensive as it is and tailoring system interactions to each user will only add to this workload. We propose machine learning algorithm-based personalized content selection to address a part of this problem and present a process for creating personalized designs that allows automating a part of the implementation. The process is based on Deterding's 2015 framework for gameful design, the lens of intrinsic skill atoms, with additional steps for selecting a personalization strategy and algorithm creation. We then demonstrate the process by implementing personalized gamification for a computer-supported collaborative learning environment. For this demonstration, we use the gamification user type hexad for personalization and the heuristics for effective design of gamification for overall design. The result of the applied design process is a context-aware, personalized gamification ruleset for collaborative environments. Lastly, we present a method for translating gamification rulesets to machine-readable classifier algorithm using the CN2 rule inducer.

Adaptive Gamification in Education: A Literature Review of Current Trends and Developments

“…They present what the engine takes into account, what it adapts, and how it adapts it. Kuntas et al [22] describe their process for designing an algorithm based personalised gamification system. They detail learner characteristics on which they base the adaptation of some game elements and the algorithm used to link the two.…”

“…Four papers from the same authors [25,28,29,30] use the term "game features" to present the same level of implementation. Knutas et al [21,22] use the terms "game like elements". Mora et al [31] present different gamification "situations" (that combine different game elements).…”

“…This is a straight forward way of implementing dynamic adaptive gamification using static adaptation rules. The systems proposed by Knutas et al [21,22] use an algorithm that also uses learners' profile and interactions. In both systems, they use the Hexad player profile, and in the more recent one [22] they also use learner skills.…”

“…The systems proposed by Knutas et al [21,22] use an algorithm that also uses learners' profile and interactions. In both systems, they use the Hexad player profile, and in the more recent one [22] they also use learner skills. In [21] they analysed videos of students during project meetings and classified their interactions and propose different game elements based on a combination of profile and interaction types.…”

Special Issue on Advances in Human-Computer Interaction

“…Creating good gamified designs is effort intensive per se. The paper "A Process for Designing Algorithm-Based Personalized Gamification Demonstrated for a Computer-Supported Collaborative Learning Environment" [6] builds on the observation that, given the demonstrated importance of taking personal characteristics and preferences into account when designing gamified systems, "tailoring system interactions to each user will only add to this workload". Hence, the authors propose to partially address this problem by producing personalized contents with the aid of appropriate machine learning algorithms.…”