AutoECART: Automatic energy conservation analysis of rovibronic transitions. (September 2021)

• Record Type: Journal Article
• Authors: Tóbiás, Roland; Bérczi, Kristóf; Szabó, Csaba; Császár, Attila G.
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 272(2021)
• Publication Date: 2021-09
• Subjects: Network theory -- Constrained empirical energies -- autoECART -- GEISA; Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2021.107756
• Languages: English
• ISSNs: 0022-4073

Abstract

Highlights: A novel network-based heuristic algorithm is proposed for the automatic detection of outliers in spectroscopic databases. A rigorous proof is provided for the necessary and sufficient condition of network consistency. Incorrect lines are identified in the GEISA-2019-H2 18O line list. Abstract: Despite decades of diligent work on their development, line-by-line (LBL) spectroscopic information systems, widely employed by scientists and engineers, may still contain a number of incorrect rovibronic lines. A novel heuristic protocol, relying on cycle bases of spectroscopic networks (SN) and a system of linear constraints, is proposed to unravel incorrect transitions present in the database. The algorithm is named autoECART, standing for automatic Energy Conservation Analysis of Rovibronic Transitions. The autoECART method is tested on synthetic SNs constructed from spectroscopic databases of nine water isotopologues. The systematic numerical tests demonstrate the outstanding capability of the autoECART procedure to identify almost all of the outliers generated randomly in these synthetic SNs. As a 'real-life' example, the GEISA-2019-H 2 18 O database is scrutinized, revealing 17 rough outlier lines. Developers of spectroscopic information systems are encouraged to utilize autoECART for cleansing the huge number of transitions deposited in LBL databases.