Minimum Entropy Indicator for Evaluating Dispersion Measure

TL;DR

This paper introduces a minimum entropy indicator based on Shannon's entropy to accurately determine dispersion measures in pulsar signals, providing a simple and effective tool for pulse profile analysis.

Contribution

It proposes a novel entropy-based indicator for dispersion measure estimation, validated with real pulsar data, improving accuracy with low computational complexity.

Findings

Minimum entropy corresponds to the true dispersion measure in most cases.

Differences with reference DMs are less than 0.5% for 45 out of 48 pulsars.

The indicator is simple to compute and effective for pulse profile research.

Abstract

Advanced radio telescopes such as the Five-hundred-meter Aperture Spherical radio Telescope (FAST) can provide high-sensitivity and high-time-resolution data of a large number of radio sources, offering an excellent opportunity for studying radio pulse profiles. However, studying pulse profiles requires the analysis of dispersion measurement (DM). The fitting method tends to make the profile conform to the model, so the fitting method is not suitable for pulse profile research. Indicators are needed to determine the profile closest to the real one, which is the profile discrimination method. This work is based on the definition of Shannon's information entropy, and believes that the pulse profile when the entropy is minimized is the closest to the true profile. This indicator is simple to calculate and can provide help for pulse profile research. This work uses real data from 48 pulsars. By calculating the information entropy of the pulsar profiles under different DMs, the DM corresponding to the minimum entropy is found, thus verifying the validity of the minimum entropy indicator. In terms of the analysis results for the 48 pulsars, the differences with the references are less than 0.5\% for all except 3 stars, and the results for these 3 stars are consistent with older references. The minimum entropy indicator can effectively obtain the DMs of radio pulse signals with low computational complexity, but it cannot be proven to be the optimal criterion. It is suggested to use multiple indicators separately when studying pulse profiles. It can be expected that the optimal indicator can provide information on the radiation mechanism of radio sources.