Dark Mode
Survey of Human-Generated Random Numbers
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This collection of data was assembled to explore the extent of human-generated randomness, addressing the core question: Can Artificial Intelligence (AI) identify hidden patterns within human choices intended to be random? Participants were instructed to select 20 random numbers, resulting in sequences that serve as a valuable resource for testing advanced predictive models, such as Long Short-Term Memory (LSTM) networks, against inherent human behavioural bias. The dataset is derived from the project "Human Randomness and AI Predictability."
Columns
The dataset contains a total of 21 columns:
- Timestamp: Records the date and time when each entry was submitted.
- R1 through R20: These 20 sequential columns capture the random numbers provided by the human contributors. Statistical analysis of these columns shows that the numerical values range from 1 to 10.
Distribution
The structure is tabular data, typically available as a CSV file named
survey.csv. It is a small dataset, with a physical size of 3.14 kB. The file is highly reliable, consisting of 50 valid records with 100% validity across all numerical fields, indicating zero missing or mismatched entries. The dataset is static, with the expected update frequency being never.Usage
This data is ideal for various analytical challenges, particularly those focused on machine learning and cognitive science:
- Testing AI Predictability: Using Regression or LSTM models to determine if patterns can be predicted within the set of human's supposedly random 20 numbers.
- Statistical Analysis: Studying the mean, standard deviation, and quantiles of human-generated number sequences to quantify deviation from true statistical randomness.
- Intermediate Data Science Projects: Suitable for educational use and projects requiring data cleaning validation and predictive modelling deployment.
Coverage
The data collection occurred over a brief period in 2021, spanning from 22 April to 4 May. There are no specific demographic or geographical details available regarding the contributors.
License
CC0: Public Domain
Who Can Use It
- Artificial Intelligence/Machine Learning Engineers: To train and evaluate algorithms specifically designed to detect subtle structure in non-random human behaviour.
- Academics and Behavioural Economists: For studying cognitive biases in decision-making and the limits of human ability to simulate randomness.
- Students and Analysts: For intermediate-level projects focused on data distribution and algorithmic prediction.
Dataset Name Suggestions
- Human Randomness Sequences for AI Prediction
- Pattern Detection in Human Pseudo-Random Choices
- Survey of Human-Generated Random Numbers
- AI Challenge: Predicting Human Intent
Attributes
Original Data Source: Survey of Human-Generated Random Numbers
QUALITYLoading...