Speaker
Mr
Ziming Liu
(Peking University)
Description
We apply principal component analysis(PCA) to simulated data of relativistic heavy-ion collisions. Unlike traditional Fourier methods, we apply PCA directly to single particle distribution. Interesting patterns are identified by PCA as eigenmodes, from which we define new flow observables $v_n^{'}$ compared to traditional ones $v_n$. The eigenmodes are very much like traditional Fourier bases, but are slightly different. Further research shows that $v_n^{'}$ are mutually more independent than $v_n$. We then relate $v_n^{'}$ to initial eccentricity $\varepsilon_n$, finding $v_n^{'}$ do have more linearity with $\varepsilon_n$ than $v_n$ with $\varepsilon_n$. This might be a signature that relativistic hydrodynamics is not as non-linear as we originally thought. With new bases chosen by PCA, the correlations between different harmonics drop significantly.
| Type | Parallel talk |
|---|---|
| Sessions (parallel only) | Heavy Ions |
Primary author
Mr
Ziming Liu
(Peking University)
Co-authors
Prof.
Huichao Song
(Peking University)
Wenbin Zhao
(school of physics Peking Uniersity)
