Back to datasets
Dataset assetOpen Source CommunitySudokuGame Difficulty Assessment

Sudoku Puzzles with Difficulty Metrics by Human Players

The dataset contains difficulty ratings of Sudoku puzzles experienced by human players, intended for evaluating Sudoku difficulty‑rating algorithms. It is built from human gameplay time and success rate, providing two difficulty metrics: $D_{TO}$ (time‑only) and $D_{TR}$ (time and completion rate).

Source
github
Created
Apr 11, 2024
Updated
May 1, 2024
Signals
98 views
Availability
Linked source ready
Overview

Dataset description and usage context

Dataset Overview

Dataset Name

A Dataset of Sudoku Puzzles with Difficulty Metrics by Human Players

Dataset Purpose

To construct a dataset containing human‑experienced difficulty metrics for Sudoku puzzles, facilitating evaluation of Sudoku difficulty‑rating algorithms.

Dataset Characteristics

  • Based on actual game records from human players.
  • Includes two difficulty metrics:
    • $D_{TO}$: considers solving time only.
    • $D_{TR}$: considers both solving time and completion rate.

Data Collection Method

Data were collected via an Android application named Cloud Sudoku.

Related Publications

  • Y.C. Lin, W.Y. Kuo, and S.W. Wang, “Cloud Sudoku: A Sudoku Application to Collect Game Records by Human Players,” in IEEE International Conference On Consumer Electronics – Taiwan (IEEE ICCE‑TW 2024), Taichung, Taiwan, July 9‑11, 2024.
Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio