Dark Mode
Bank Note Authentication Feature Set
Finance & Banking Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Features were extracted from images of laboratory specimens, including both genuine and forged banknote examples. This data is highly valuable for building robust classification models aimed at currency verification and fraud detection. For digitisation, an industrial camera typically used for print inspection captured the source images, which were 400x400 pixel gray-scale visuals achieved at a resolution of roughly 660 dpi. The features themselves were subsequently processed and derived using the Wavelet Transform tool.
Columns
This dataset contains five primary columns:
variance: A continuous value representing the variance derived from the Wavelet Transformed image.skewness: A continuous value indicating the skewness of the Wavelet Transformed image.curtosis: A continuous value detailing the curtosis of the Wavelet Transformed image.entropy: A continuous value quantifying the image entropy.class: An integer label defining the banknote's status (0 for genuine, 1 for forged).
Distribution
The data is stored in the file
BankNote_Authentication.csv, which has a file size of 46.44 kB. It is structured with 5 columns and comprises 1372 valid records in total. A key attribute of this data is its cleanliness, as there are zero missing or mismatched values across all records. The dataset is static, with an expected update frequency of Never.Usage
This feature set is ideally suited for a variety of analytical and machine learning tasks:
- Developing binary classification algorithms to accurately distinguish between authentic and counterfeit currency.
- Conducting Data Analytics to explore the relationships and importance of wavelet-derived features.
- Supporting Data Visualization efforts to map the distribution characteristics of the extracted metrics.
- Practical exercises in data cleaning and pre-processing, although the data provided is notably clean.
Coverage
The data is strictly derived from the image characteristics of the banknote-like specimens themselves. It carries no geographic, temporal, or human demographic scope.
License
CC0: Public Domain
Who Can Use It
The primary audience includes:
- Machine Learning Engineers: For training and evaluating classification models on a high-quality, benchmark dataset.
- Data Scientists: For statistical analysis of image properties and feature engineering studies.
- Academics and Students: As a clean and reliable source for learning and testing fundamental classification techniques.
- Banking Technology Researchers: Individuals interested in automated systems for currency verification and financial fraud detection.
Dataset Name Suggestions
- Bank Note Authentication Feature Set
- Wavelet Banknote Image Attributes
- Currency Forgery Detection Features
- UCI Banknote Wavelet Data
Attributes
Original Data Source: Bank Note Authentication Feature Set
QUALITYLoading...