arXiv subjects

Eric Endress

Publications and source records attributed to Eric Endress.

6 recordsLinked to original sources

arXiv2024

Forecasting Company Fundamentals

Company fundamentals are key to assessing companies' financial and overall success and stability. Forecasting them is important in multiple fields, including investing and econometrics. While statistical and contemporary machine learning methods have been applied to many time series tasks, there is a lack of comparison of these approaches on this particularly challenging data regime. To this end, we try to bridge this gap and thoroughly evaluate the theoretical properties and practical performance of 24 deterministic and probabilistic company fundamentals forecasting models on real company data. We observe that deep learning models provide superior forecasting performance to classical models, in particular when considering uncertainty estimation. To validate the findings, we compare them to human analyst expectations and find that their accuracy is comparable to the automatic forecasts. We further show how these high-quality forecasts can benefit automated stock allocation. We close by presenting possible ways of integrating domain experts to further improve performance and increase reliability.

q-fin.ST↗

arXiv2017

Testing FLUKA on neutron activation of Si and Ge at nuclear research reactor using gamma spectroscopy

Samples of two characteristic semiconductor sensor materials, silicon and germanium, have been irradiated with neutrons produced at the RP-10 Nuclear Research Reactor at 4.5 MW. Their radionuclides photon spectra have been measured with high resolution gamma spectroscopy, quantifying four radioisotopes ($^{28}$Al, $^{29}$Al for Si and $^{75}$Ge and $^{77}$Ge for Ge). We have compared the radionuclides production and their emission spectrum data with Monte Carlo simulation results from FLUKA. Thus we have tested FLUKA's low energy neutron library (ENDF/B-VIIR) and decay photon scoring with respect to the activation of these semiconductors. We conclude that FLUKA is capable of predicting relative photon peak amplitudes, with gamma intensities greater than 1%, of produced radionuclides with an average uncertainty of 13%. This work allows us to estimate the corresponding systematic error on neutron activation simulation studies of these sensor materials.

physics.ins-det↗

arXiv2016

Variance reduction with practical all-to-all lattice propagators

We discuss all-to-all quark propagator techniques in two (related) contexts within Lattice QCD: the computation of closed quark propagators, and applications to the so-called "eye diagrams" appearing in the computation of non-leptonic kaon decay amplitudes. Combinations of low-mode averaging and diluted stochastic volume sources that yield optimal signal-to-noise ratios for the latter problem are developed. We also apply a recently proposed probing algorithm to compute directly the diagonal of the inverse Dirac operator, and compare its performance with that of stochastic methods. At fixed computational cost the two procedures yield comparable signal-to-noise ratios, but probing has practical advantages which make it a promising tool for a wide range of applications in Lattice QCD.

hep-lat↗

arXiv2016

Exploring the role of the charm quark in the $ΔI=1/2$ rule

We study the dependence on the charm quark mass of the leading-order low-energy constants of the $ΔS=1$ effective Hamiltonian, with the aim of elucidating the role of the charm mass scale in the $ΔI=1/2$ rule for $K\toππ$ decay. To that purpose, finite-volume Chiral Perturbation Theory predictions are matched to QCD simulations, performed in the quenched approximation with overlap fermions and $m_u=m_d=m_s$. Light quark masses range between a few MeV up to around one third of the physical strange mass, while charm masses range between $m_u$ and a few hundred MeV. Novel variance reduction techniques are used to obtain a signal for penguin contractions in correlation functions involving four-fermion operators. The important role played by the subtractions required to construct renormalised amplitudes for $m_c \neq m_u$ is discussed in detail. We find evidence that the moderate enhancement of the $ΔI=1/2$ amplitude previously found in the GIM limit $m_c=m_u$ increases only slightly as $m_c$ abandons the light quark regime. Hints of a stronger enhancement for even higher values of $m_c$ are also found, but their confirmation requires a better understanding of the subtraction terms.

hep-lat↗

arXiv2012

Contribution of the charm quark to the ΔI=1/2 rule

We report on the progress of our ongoing project to quantify the role of the charm quark in the non-leptonic decay of a kaon into two pions. The effect of its associated mass scale in the dynamics underlying the ΔI = 1/2 rule can be studied by monitoring the dependence of kaon decay amplitudes on the charm quark mass using an effective ΔS = 1 weak Hamiltonian. In contrast to commonly used approaches the charm quark is kept as an active degree of freedom. Quenched results in the GIM limit have shown that a significant part of the ΔI = 1/2 enhancement is purely due to low-energy QCD effects. Moving away from the GIM limit involves the computation of diagrams containing closed quark loops which requires new variance reduction techniques in order to determine the relevant weak effective low-energy couplings. We employ a combination of low-mode averaging and stochastic volume sources in order to compute these diagrams and observe a significant improvement in the statistical signal.

hep-lat↗

arXiv2012

Variance reduction techniques for a quantitative understanding of the ΔI = 1/2 rule

The role of the charm quark in the dynamics underlying the ΔI = 1/2 rule for kaon decays can be understood by studying the dependence of kaon decay amplitudes on the charm quark mass using an effective ΔS = 1 weak Hamiltonian in which the charm is kept as an active degree of freedom. Overlap fermions are employed in order to avoid renormalization problems, as well as to allow access to the deep chiral regime. Quenched results in the GIM limit have shown that a significant part of the enhancement is purely due to low-energy QCD effects; variance reduction techniques based on low-mode averaging were instrumental in determining the relevant weak effective lowenergy couplings in this case. Moving away from the GIM limit requires the computation of diagrams containing closed quark loops. We report on our progress to employ a combination of low-mode averaging and stochastic volume sources in order to control these contributions. Results showing a significant improvement in the statistical signal are presented.

hep-lat↗