遇见数据集

LAMOST DR11 Spectra Dataset and DESI DR1 Carbon-Star Candidate Catalogs

收藏
Zenodo2026-09-01 更新2026-10-01 收录
官方服务:

资源简介:

This repository provides the datasets and catalogs associated with our study of carbon-star identification using LAMOST DR11 and DESI DR1 spectroscopy. The repository contains the following data products: LAMOST DR11 Spectra Dataset This dataset contains LAMOST DR11 optical spectra used for the carbon-star classification experiments presented in our study. It comprises spectra from seven source classes: carbon stars, F-, G-, K-, and M-type stars, galaxies, and QSOs. For model training and evaluation, these seven classes are formulated as a binary classification task, with carbon stars treated as the positive class and the remaining six classes combined as the non-carbon class. All training, validation, and test partitions were constructed at the object level to prevent spectra of the same astronomical source from appearing in different subsets. The spectral data are provided in FITS format. The dataset includes: Training set Validation set Test set DESI DR1 High-Confidence Carbon-Star Candidate Catalog This catalog contains high-confidence carbon-star candidates identified from DESI DR1 spectra. For each candidate, the catalog provides the DESI TARGETID, celestial coordinates, Gaia source ID, signal-to-noise ratios in the DESI B, R, and Z arms, pseudo-equivalent widths (pEWs) of the C₂ and CN molecular bands, coadded exposure time, optical and infrared photometric fluxes and their inverse variances, g- and r-band AB magnitudes, synthetic i-band AB magnitudes derived from the DESI spectra, Gaia G, BP, and RP magnitudes, and good-pixel fractions for the DESI B, R, and Z arms. DESI DR1 Moderate-Confidence Carbon-Star Candidate Catalog This catalog contains moderate-confidence carbon-star candidates identified from DESI DR1 spectra. The catalog provides the same parameters as the high-confidence carbon-star candidate catalog. The source code for data preprocessing, model training, candidate identification, and evaluation is available from the accompanying GitHub repository: https://github.com/yangshengzhen327/Carbon-Stars

提供机构:
Zenodo
创建时间:
2026-09-01
二维码
社区交流群
二维码
科研交流群
商业服务