home.social

#statistical — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #statistical, aggregated by home.social.

  1. #paperOfTheDay "Relations between short-range and long-range Ising models" from 2014.
    The basic version of the Ising model in #statistical #physics is a lattice where every site contains a binary variable, a "spin" that can point up or down. There is a nearest-neighbour interaction which energetically prefers neighbouring spins to point in the same direction. Then, there are "long range" versions, where the interaction also takes into account spins at larger distance, with a weighting factor that decays with some power law with exponent sigma (where sigma=2 reproduces the conventional short-range model). In this class of models, there are thus two parameters: The dimension d, and the parameter sigma.
    The present paper is mostly a numerical Monte Carlo study of various such systems at different d and sigma. The guiding question is whether instead of two, there is actually only one parameter. Phrased differently: Given some d and sigma, can I find some other D such that the short-range (sigma=2) model at this D is equivalent to the long-range one at d? It is intuitively plausible that this works close to the interface between long-range and short-range models, making it slightly long-range is mostly the same as slightly changing dimension. However, the authors demonstrate that such relations are in fact only an approximation, and fail for generic values of sigma and d.
    A second finding concerns the structure of the correlation functions of certain long-range models, which decay according to a power law (as expected from the #renormalization group), but where the leading correction is another power law (and not exponentially small).
    journals.aps.org/pre/abstract/
    arxiv.org/abs/1401.6805

  2. I'll just put this here for the #record in case anyone wants to #question me about it later.

    I #donated with #BitCoin back in January to #AnnasArchive and now the site's knocked #offline literally.

    I paid like C$120.00 for the online form of the #project just to disappear.

    Yeah ok #wikipedia has them moving #domains every other day but it was a lot more stable than that before I gave them a leg up.

    This is why I hate centralized #websites and hosting.

    This is why #self #hosting works by connecting #dedicated #servers to the #internet as a whole.

    If you use anything less, you will face more #legal challenges hosting #human #knowledge and unfortunately, it is as #simple as that.

    You cannot convince me that #BigTech is not turning into #authorized #tech as it seems as though the #copyright #industry and the #torent industry has eluded to colluding with one another to obtain #rights for.

    Guys, I don't know what to say anymore. I'm a huge fan of #books and #media and #free #open #source #software and I long for the day that the #government doesn't need to get a cut of everything I #download or #upload anywhere, especially for #personal #use over all.

    Please do whatever you can to fight #statistical #bottlenecks where ever they appear. I want humanity to have better access to #education and #intelligence built by real people in a fair and consistent manner.

    I always am striving and trying my best to be a better #human being.

  3. Your #process #data holds early warnings—#LabPlot (#free, #opensource) turns them into clear #statistical alerts, so you act before issues escalate.

    Unlike complex tools, LabPlot's #StatisticalProcessControl tools are designed for quick, reliable #insights.

    Using #SPC, reductions in the process #variation of 50% are not unusual.

    Try it now:
    1. Download: labplot.org/download.
    2, Add a #ControlChart.
    3. Act on #facts.

    #DataAnalysis #DataViz #Statistics #Business #Chemistry #Physics #Health

  4. appears to be an uncanny fit (orange) to 400 measured values (blue) using 16 sinusoidal frequencies! this time it's grant work, not contract work, so i will be able to explain this in more detail eventually... for now, it's just a work in progress that looks cool :)

    #dataScience #dataAnalysis #spectralAnalysis #leastSquares #OLS #sinusoidal #trigonometric #regression #modeling #statistical

  5. Detective work with #genomes: some invasive species have been introduced deliberately, others inadvertently. A new #statistical framework helps to track what probably happened. Demonstration object is the Pacific oyster. #bayes pnas.org/doi/abs/10.1073/pnas.

  6. The study of topological phases in quantum systems (like in the #quantum #Hall effect) can be seen as analogous to the study of #topological #transitions in #statistical models like the #Ising #model.

  7. New paper. A total of about 10 years of work.

    Approximately 40% of all proteins are intrinsically disordered or contain intrinsically disordered regions. So far, it was widely believed that these regions always had a random coil secondary structure. We show that this is not necessarily the case.

    We applied the replica method to a generalized Langevin equation to describe the structure of AXR3, which is a co-receptor of the plant growth hormone auxin, and found that it adopts a number of distinct, different tertiary structures. Of these, two are particularly stable, almost equally probable, and account for 90% of the conformations. The results are confirmed by NMR spectroscopy and circular dichroism experiments. This shows that intrinsic disorder does not always mean a constantly disordered structure, such as a random coil or a Gaussian self-avoiding walk. Rather, disorder can arise as an ensemble effect of a glassy energy landscape, in which a protein can exist in a mixture of separate folded structures, with an energy barrier that may also induce symmetry breaking and the effective loss of ergodicity, depending on its height.

    The fact that the first fully characterized example of such cases is a plant protein from the auxin system is an added bonus. 😎

    #computationalbiology #biology #biophysics #statistical #physics #Complexity #ComplexSystems #plantbiology #hormones #auxin #moleculardynamics #simulation #NMR #circulardichroism #structure #disorder #IDP #IDR

    pnas.org/doi/10.1073/pnas.2221

  8. Understanding ‘it depends’ in #ecology: a guide to #hypothesising, #visualising and #interpreting #statistical #interactions

    #OpenAccess
    #BiologicalReviews

    "Two fundamental properties of interactions are often overlooked during the process of analysing interactions between drivers: the measurement scale and the symmetry. Overlooking these properties can lead to three inferential errors"

    onlinelibrary.wiley.com/doi/fu

  9. arxiv.org/abs/2101.12223

    I'm city hopping by #train from Warsaw to #Benasque and yesterday I met Oli Reardon-Smith in Kraków, here a paper with Hakop, Kamil and Stephen
    arxiv.org/abs/2101.12223
    I learned about an innovation in that they use part of samples to estimate certain parameters needed for a large deviation bound. It's like #Hoefding's inequality on wheels.

    I know of the #BCa (biased corrected accelerated) #bootstrap method for assigning confidence intervals to #statistical estimations. Here a round of resampling is used to estimate two parameters to correct bias and accelerate sample efficiency in pinning down the accurate confidence interval.

    What are other methods where you sample to get information about what to sample subsequently?