arXiv2019
We present the determination of stellar parameters and individual elemental abundances for 6 million stars from $\sim$8 million low-resolution ($R\sim1800$) spectra from LAMOST DR5. This is based on a modeling approach that we dub $The$ $Data$--$Driven$ $Payne$ ($DD$--$Payne$), which inherits essential ingredients from both {\it The Payne} \citep{Ting2019} and $The$ $Cannon$ \citep{Ness2015}. It is a data-driven model that incorporates constraints from theoretical spectral models to ensure the derived abundance estimates are physically sensible. Stars in LAMOST DR5 that are in common with either GALAH DR2 or APOGEE DR14 are used to train a model that delivers stellar parameters ($T_{\rm eff}$, $\log g$, $V_{\rm mic}$) and abundances for 16 elements (C, N, O, Na, Mg, Al, Si, Ca, Ti, Cr, Mn, Fe, Co, Ni, Cu, and Ba) when applied to LAMOST spectra. Cross-validation and repeat observations suggest that, for ${\rm S/N}_{\rm pix}\ge 50$, the typical internal abundance precision is 0.03--0.1\,dex for the majority of these elements, with 0.2--0.3\,dex for Cu and Ba, and the internal precision of $T_{\rm eff}$ and $\log g$ is better than 30\,K and 0.07\,dex, respectively. Abundance systematics at the $\sim$0.1\,dex level are present in these estimates, but are inherited from the high-resolution surveys' training labels. For some elements, GALAH provides more robust training labels, for others, APOGEE. We provide flags to guide the quality of the label determination and to identify binary/multiple stars in LAMOST DR5. The abundance catalogs are publicly accessible via \href{url}{http://dr5.lamost.org/doc/vac}.