Category Archives: Bayesian inversion

On Smart Data

Posted on 11 June 2016 by ecoquant

One of the things I find surprising, if not astonishing, is that in the rush to embrace Big Data, a lot of learning and statistical technique has been left apparently discarded along the way. I’m hardly the first to point … Continue reading →

Posted in Akaike Information Criterion, Bayes, Bayesian, Bayesian inversion, big data, bigmemory package for R, changepoint detection, data science, data streams, dlm package, dynamic generalized linear models, dynamic linear models, dynamical systems, Generalize Additive Models, generalized linear models, information theoretic statistics, Kalman filter, linear algebra, logistic regression, machine learning, Markov Chain Monte Carlo, mathematics, mathematics education, maths, maximum likelihood, MCMC, Monte Carlo Statistical Methods, multivariate statistics, numerical analysis, numerical software, numerics, quantitative biology, quantitative ecology, rationality, reasonableness, sampling, smart data, state-space models, statistical dependence, statistics, the right to know, time series | Leave a comment

p-values and hypothesis tests: the Bayesian(s) rule

Posted on 7 March 2016 by ecoquant

The American Statistical Association of which I am a longtime member issued an important statement today which will hopefully move statistical practice in engineering and especially in the sciences away from the misleading practice of using p-values and hypothesis tests. … Continue reading →

Posted in approximate Bayesian computation, arXiv, Bayes, Bayesian, Bayesian inversion, bollocks, Christian Robert, climate, complex systems, data science, Frequentist, information theoretic statistics, likelihood-free, Markov Chain Monte Carlo, MCMC, Monte Carlo Statistical Methods, population biology, rationality, reasonableness, science, scientific publishing, statistical dependence, statistics, stochastics, Student t distribution | Leave a comment

“Grid shading by simulated annealing” [Martyn Plummer]

Posted on 9 February 2016 by ecoquant

Source: Grid shading by simulated annealing (or what I did on my holidays), aka “fun with GCHQ job adverts”, by Martyn Plummer, developer of JAGS. Excerpt: I wanted to solve the puzzle but did not want to sit down with … Continue reading →

Posted in approximate Bayesian computation, Bayesian, Bayesian inversion, Boltzmann, BUGS, Christian Robert, Gibbs Sampling, JAGS, likelihood-free, Markov Chain Monte Carlo, Martyn Plummer, mathematics, maths, MCMC, Monte Carlo Statistical Methods, optimization, probabilistic programming, SPSA, stochastic algorithms, stochastic search | Leave a comment

high dimension Metropolis-Hastings algorithms

Posted on 6 February 2016 by ecoquant

If attempting to simulate from a multivariate standard normal distribution in a large dimension, when starting from the mode of the target, i.e., its mean γ, leaving the mode γis extremely unlikely, given the huge drop between the value of the density at the mode γ and at likely realisations Continue reading →

Posted in Bayes, Bayesian, Bayesian inversion, boosting, chance, Christian Robert, computation, ensembles, Gibbs Sampling, James Spall, Jerome Friedman, Markov Chain Monte Carlo, mathematics, maths, MCMC, Monte Carlo Statistical Methods, multivariate statistics, numerical software, numerics, optimization, reasonableness, Robert Schapire, SPSA, state-space models, statistics, stochastic algorithms, stochastic search, stochastics, Yoav Freund | Leave a comment

Generating supports for classification rules in black box regression models

Posted on 25 January 2016 by ecoquant

Inspired by the extensive and excellent work in approximate Bayesian computation (see also), especially that done by Professors Christian Robert and colleagues (see also), and Professor Simon Wood (see also), it occurred to me that the complaints regarding lack of … Continue reading →

Posted in approximate Bayesian computation, Bayes, Bayesian, Bayesian inversion, generalized linear models, machine learning, numerical analysis, numerical software, probabilistic programming, rationality, reasonableness, state-space models, statistics, stochastic algorithms, stochastic search, stochastics, support of black boxes | Leave a comment

reblog: “Tiny Data, Approximate Bayesian Computation and the Socks of Karl Broman”

Posted on 24 November 2015 by ecoquant

It’s Rasmus Bååth, in a post and video of which I am very fond: http://www.sumsar.net/blog/2014/10/tiny-data-and-the-socks-of-karl-broman/.

Posted in approximate Bayesian computation, Bayesian, Bayesian inversion, empirical likelihood, evidence, likelihood-free, probability, rationality, reasonableness, statistics, stochastic algorithms, stochastic search | 1 Comment

On differential localization of tumors using relative concentrations of ctDNA. Part 2.

Posted on 26 October 2015 by ecoquant

Part 1 of this series introduced the idea of ctDNA and its use for detecting cancers or their resurgence, and proposed a scheme whereby relative concentrations of ctDNA at two or more sites after controlled disturbance might be used to … Continue reading →

Posted in Bayes, Bayesian, Bayesian inversion, cancer research, ctDNA, differential equations, diffusion, diffusion processes, engineering, linear algebra | 1 Comment

On differential localization of tumors using relative concentrations of ctDNA. Part 1.

Posted on 11 October 2015 by ecoquant

Like most mammalian tissue, tumors often produce shards of DNA as a byproduct of cell death and fracture. This circulating tumor DNA is being studied as a means of detecting tumors or their resurgence after treatment. (See also a Q&A … Continue reading →

Posted in approximate Bayesian computation, Bayesian, Bayesian inversion, cardiovascular system, diffusion, dynamic linear models, eigenanalysis, engineering, forecasting, mathematics, maths, medicine, networks, prediction, spatial statistics, statistics, stochastic algorithms, stochastic search, wave equations | 3 Comments