Research Resources

Datasets

Access and explore our research data collections.

Polynomials’ Roots Dataset

Our Polynomials' Roots dataset is a comprehensive resource for developing and validating root-finding algorithms. The files are saved in CSV format, with coefficients denoted by a_i and roots identified by alpha_j (or re_alpha_j/im_alpha_j for complex cases).

Data was generated using Mathematica software with double-precision arithmetic.

Context

Finding the arbitrary roots (real or complex) of a given polynomial is a fundamental task in various areas of science and engineering. Most available methods lead to accumulation of rounding errors and inaccurate results. This dataset is designed for testing and comparing tools that compute the roots of polynomials, particularly artificial intelligence solutions.

Content Structure

Two main directories: "real" (polynomials with only real roots) and "real_complex" (polynomials with both real and complex roots). These store coefficients (deg_n_coef) and roots (deg_n_roots) for univariate polynomials of degrees 5, 10, 15, 20 and 25.

Attribute Information

Roots were generated in the interval [-1, 1] for real-only polynomials. For real/complex polynomials, coefficients were generated in the interval [0, 1]. The number of attributes is equal to n (real) or n*2 (real/complex).

General Information

100,000 instances per polynomial degree. Recommended task: Regression. There are no missing values.

Citation Request

If you use this dataset, please cite: "A Neural Network-Based Approach for Approximating the Arbitrary Roots of Polynomials", IEEE Access.

Other Datasets

Soon we will release other datasets regarding Covid-19 and Football.

To Request Data

Fill the form below to request access to our datasets. Please identify your institution and the planned usage.