Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Explainable e-sports win prediction through Machine Learning classification in streaming

Silvia García-Méndez, Francisco de Arriba-Pérez

TL;DR

The paper tackles real-time e-sports win prediction by integrating streaming machine learning with sliding-window data controls to capture evolving game dynamics, specifically in CS:GO. It introduces a four-part system (in-game data fusion, stream-based processing, stream-based classification, and post-game explainability) and demonstrates an online classifier pipeline that achieves >90% accuracy while providing natural-language explanations through an explainability dashboard. Key contributions include a sliding-window feature engineering scheme, variance-threshold feature selection tailored to streaming, and an interpretable post-hoc explanation mechanism suitable for ranking and recommender systems. The work has practical impact by enabling trust-aware, real-time decision support in e-sports analytics and paving the way for live API integration and broader production deployment.

Abstract

The increasing number of spectators and players in e-sports, along with the development of optimized communication solutions and cloud computing technology, has motivated the constant growth of the online game industry. Even though Artificial Intelligence-based solutions for e-sports analytics are traditionally defined as extracting meaningful patterns from related data and visualizing them to enhance decision-making, most of the effort in professional winning prediction has been focused on the classification aspect from a batch perspective, also leaving aside the visualization techniques. Consequently, this work contributes to an explainable win prediction classification solution in streaming in which input data is controlled over several sliding windows to reflect relevant game changes. Experimental results attained an accuracy higher than 90 %, surpassing the performance of competing solutions in the literature. Ultimately, our system can be leveraged by ranking and recommender systems for informed decision-making, thanks to the explainability module, which fosters trust in the outcome predictions.

Explainable e-sports win prediction through Machine Learning classification in streaming

TL;DR

The paper tackles real-time e-sports win prediction by integrating streaming machine learning with sliding-window data controls to capture evolving game dynamics, specifically in CS:GO. It introduces a four-part system (in-game data fusion, stream-based processing, stream-based classification, and post-game explainability) and demonstrates an online classifier pipeline that achieves >90% accuracy while providing natural-language explanations through an explainability dashboard. Key contributions include a sliding-window feature engineering scheme, variance-threshold feature selection tailored to streaming, and an interpretable post-hoc explanation mechanism suitable for ranking and recommender systems. The work has practical impact by enabling trust-aware, real-time decision support in e-sports analytics and paving the way for live API integration and broader production deployment.

Abstract

The increasing number of spectators and players in e-sports, along with the development of optimized communication solutions and cloud computing technology, has motivated the constant growth of the online game industry. Even though Artificial Intelligence-based solutions for e-sports analytics are traditionally defined as extracting meaningful patterns from related data and visualizing them to enhance decision-making, most of the effort in professional winning prediction has been focused on the classification aspect from a batch perspective, also leaving aside the visualization techniques. Consequently, this work contributes to an explainable win prediction classification solution in streaming in which input data is controlled over several sliding windows to reflect relevant game changes. Experimental results attained an accuracy higher than 90 %, surpassing the performance of competing solutions in the literature. Ultimately, our system can be leveraged by ranking and recommender systems for informed decision-making, thanks to the explainability module, which fosters trust in the outcome predictions.
Paper Structure (18 sections, 2 equations, 7 figures, 8 tables)

This paper contains 18 sections, 2 equations, 7 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. System architecture
  4. In-game data fusion
  5. Stream-based data processing
  6. Pre-game feature engineering
  7. Feature analysis & selection
  8. Stream-based classification
  9. Post-game stream-based explainability
  10. Evaluation and discussion
  11. Experimental dataset
  12. In-game data fusion
  13. Stream-based data processing
  14. Pre-game feature engineering
  15. Feature analysis & selection
  16. ...and 3 more sections

Figures (7)

  • Figure 1: System diagram composed of: (i) in-game data fusion, (ii) stream-based data processing, (iii) stream-based classification, and (iv) post-game stream-based explainability.
  • Figure 2: Features engineered using the sliding windows.
  • Figure 3: Streaming and batch supervised classification.
  • Figure 4: Distribution function of the number of games per player.
  • Figure 5: Explainability dashboard exploiting post-game features and combining visualization and natural language.
  • ...and 2 more figures