Posts on Ioannis Kosmidis
http://www.ikosmidis.com/post/
Recent content in Posts on Ioannis KosmidisHugo -- gohugo.ioenWed, 14 Oct 2020 00:00:00 +0000Accessibility of HTML pages from R Markdown
http://www.ikosmidis.com/post/blog-accessibility/
Wed, 14 Oct 2020 00:00:00 +0000http://www.ikosmidis.com/post/blog-accessibility/Accessibility Over the years, I faced situations where I needed to make content (lecture notes, manuscripts, consultancy reports, etc.) as accessible as possible to effectively communicate with colleagues, clients, and students with particular accessibility requirements.
All those “accessibility encounters” of mine left me with a pile of unstructured notes and advice to my self to which I often refer. Well, I decided to blog about parts of it in case anyone else finds it useful.brquasi: Improved quasi-likelihood estimation
http://www.ikosmidis.com/post/news-brquasi-erum2020/
Wed, 10 Jun 2020 00:00:00 +0000http://www.ikosmidis.com/post/news-brquasi-erum2020/On 19 June 2020, I will be giving a talk at the eRum2020 conference titled “brquasi: Improved quasi-likelihood estimation”. The talk is about the brquasi R package, which provides a glm method for improving bias in quasi likelihood estimation.
That package is a spin-off of theoretical work with Nicola Lunardon on methods for reduced-bias M-estimation in general that require neither re-sampling (like the bootstrap does) nor full knowledge of the underlying distribution of the data (like other higher-order asymptotic methods do).detectseparation R package is on CRAN
http://www.ikosmidis.com/post/news-detectseparation-march-2020/
Thu, 26 Mar 2020 00:00:00 +0000http://www.ikosmidis.com/post/news-detectseparation-march-2020/detectseparation provides pre-fit and post-fit methods for detecting separation and infinite maximum likelihood estimates in generalized linear models with categorical responses.
Pre-fit methods The pre-fit methods apply on binomial-response generalized liner models such as logit, probit and cloglog regression, and can be directly supplied as fitting methods to the glm() function. They solve the linear programming problems for the detection of separation developed in Konis (2007), using ROI or lpSolveAPI.Making sense of CRAN: Package and collaboration networks
http://www.ikosmidis.com/post/news-cranly-user2019/
Sun, 07 Jul 2019 00:00:00 +0000http://www.ikosmidis.com/post/news-cranly-user2019/On 12 July 2019, I will be giving a talk at the useR! 2019 conference titled “Making sense of CRAN: Package and collaboration networks”. The talk is about what my cranly R package can do and how it can help users and developers understand CRAN better and the position of their (intended) packages within it.
Update (12 July 2019) The slides are available hereA workflow that most probably isn’t yours
http://www.ikosmidis.com/post/blog-workflows/
Tue, 18 Jun 2019 00:00:00 +0000http://www.ikosmidis.com/post/blog-workflows/In early 2019, I was invited to give a talk at the Young Researcher’s Meetings of the Department of Statistics at University of Warwick (the place I work). The Young Researchers’ meeting is a weekly seminar for postgraduate students and postdocs to discuss research and exchange ideas.
My immediate thought was to talk about one of my current and exciting research projects. Of course, I know from experience that the talks I have given on research projects I am excited about are not necessarily equally exciting for the audience.Location-adjusted Wald statistics for scalar parameters appears in CSDA
http://www.ikosmidis.com/post/news-lawald-paper/
Tue, 16 Apr 2019 00:00:00 +0000http://www.ikosmidis.com/post/news-lawald-paper/Mean and median bias reduction in generalized linear models appears in Statistics and Computing (joint work with Kenne Pagui E C, and Sartori N) appeared online in Statistics and Computing.trackeRapp: interface & workflow for running, cycling and swimming data
http://www.ikosmidis.com/post/news-trackerapp-is-out/
Fri, 12 Apr 2019 00:00:00 +0000http://www.ikosmidis.com/post/news-trackerapp-is-out/🏃♀️🚴♂️🏊♀️
The first version of our trackeRapp R package is out!
trackeRapp provides the first fully interactive data analysis workflow for your runs, rides and swims! It is completely open source and written in R. Its backend is based on the trackeR R package and the frontend has been developed using the shiny R package.
To launch the trackeRapp interface locally, launch R and then
install.packages("trackeRapp") trackeRapp::trackeRapp() Then, hit “Load” and “Upload sample data set” to see what {trackeRapp} can do!Mean and median bias reduction in generalized linear models appears in Statistics and Computing
http://www.ikosmidis.com/post/news-mean-median-paper/
Mon, 04 Mar 2019 00:00:00 +0000http://www.ikosmidis.com/post/news-mean-median-paper/Mean and median bias reduction in generalized linear models appears in Statistics and Computing (joint work with Kenne Pagui E C, and Sartori N) appeared online in Statistics and Computing.cranly: Package and collaboration networks in CRAN
http://www.ikosmidis.com/post/news-cranly-oxfordrug2018/
Sat, 03 Nov 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-cranly-oxfordrug2018/On 5 November 2018, I will be giving a talk at the R user group Oxford about “Package and collaboration networks in CRAN” using my cranly R package.
Update (06 Nov 2018) The slides and the code from my talk are here and here.trackeR v1.1 is on CRAN
http://www.ikosmidis.com/post/tracker-v11/
Mon, 24 Sep 2018 00:00:00 +0000http://www.ikosmidis.com/post/tracker-v11/The trackeR R package v1.1 is on CRAN. Version 1.1 of the package introduces full support for multi-sport workouts, numerous new visualisations, vast under-the-hood enhancements, bug fixes and more.
See NEWS for complete lists of new functionality, enhancements and bug fixes.
The package vignette provides a quick tour on its functionality. More detailed descriptions of the package and its methods, and real-data demonstrations of the package functionality can be found here, which is an updated version ofAn R recursion to compute Bartlett relations
http://www.ikosmidis.com/post/blog-bartlett/
Sun, 09 Sep 2018 00:00:00 +0000http://www.ikosmidis.com/post/blog-bartlett/It is one of those times that I am really excited and somewhat obsessed with the current research project I am working on. It is an old idea that after many years going around my head, it seems to be coming close to fruition. This is not the reason I am writing though.
Part of that research project involves heavy usage of the Bartlett relations (or Bartlett identities). The Bartlett relations (see Bartlett 1953, Section 2) are at the core of much of statistical theory and practice.Modelling outcomes of soccer matches appears in Machine Learning
http://www.ikosmidis.com/post/news-soccer-paper/
Wed, 01 Aug 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-soccer-paper/Modelling outcomes of soccer matches (joint work with Tsokos A, Narayanan S, Baio G, Cucuringu M, Whitaker G and Király F J) appeared in Machine learning.Median bias reduction in random-effects meta-analysis and meta-regression appears in Statistical Methods for Medical Research
http://www.ikosmidis.com/post/news-median-meta-paper/
Wed, 02 May 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-median-meta-paper/Median bias reduction in random-effects meta-analysis and meta-regression (joint work with Kyriakou S, and Sartori N) appeared online in Statistical Methods in Medical Research.Turing Data Science Masterclass: 'Scalable regression: old tricks, new era'
http://www.ikosmidis.com/post/news-turing_masterclass/
Tue, 01 May 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-turing_masterclass/On 25th May 2018, I will be giving a Turing Data Science Masterclass titled “Scalable regression: old tricks, new era”. Information and registration details can be found here.Warwick Impact Fund Award
http://www.ikosmidis.com/post/news-warwick-impacy-fund/
Fri, 27 Apr 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-warwick-impacy-fund/I received a Warwick Impact Fund Award to further the impact of my trackeR R package through an publicly accessible, open-source interface.My cranly R package is on CRAN
http://www.ikosmidis.com/post/news-cranly-march-2018/
Wed, 28 Mar 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-cranly-march-2018/cranly provides comprehensive methods for cleaning up and organising the information in the CRAN package database and for building package directives networks and collaboration networks. See also the package vignettes and the cranly GitHub page for more details.
Take a look at my Software page for the cranly directives network for my R packages and for how to make your own!My logo is a bivariate density
http://www.ikosmidis.com/post/blog-logo/
Sat, 03 Mar 2018 00:00:00 +0000http://www.ikosmidis.com/post/blog-logo/My logo in GitHub (see here) is the contours of the density function of a mixture of bivariate normal distributions (see Figure 1). The idea came to me when I was working on the paper
Kosmidis I and Karlis D (2016). Model-based clustering using copulas with applications. Statistics and Computing, 26, 1079–1099 DOI ArXiV In that paper, amongst several other things, we introduce the use of rotations of the components of mixtures of copulas to define extremely flexible mixture models.Associate Editor in Biometrika
http://www.ikosmidis.com/post/news-biometrika/
Tue, 09 Jan 2018 00:00:00 +0000http://www.ikosmidis.com/post/news-biometrika/I have joined the Editorial Board of Biometrika, serving as an Associated Editor.trackeR paper appears in JSS
http://www.ikosmidis.com/post/news-tracker-paper/
Tue, 12 Dec 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-tracker-paper/trackeR: Infrastructure for running and cycling data from GPS-enabled tracking devices in R (joint work with Hannah Frick) appeared online in the Journal of Statistical Software.Deterministic approximation methods and pseudo-likelihoods at BayesComp 2018
http://www.ikosmidis.com/post/news-bayecomp2018/
Sun, 01 Oct 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-bayecomp2018/I am organising the session “Deterministic approximation methods and pseudo-likelihoods” at BayesComp 2018 (26th - 28th March 2018, Barcelona, Spain). The session will consist of talks from Nancy Reid (Univeristy of Toronto, Canada), Helen Ogden (University of Southampton, UK) and Erlis Ruli (University of Padova, Italy).Speaking at the AUEB Sports Analytics Workshop
http://www.ikosmidis.com/post/news-aueb-sports-workshop/
Wed, 16 Aug 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-aueb-sports-workshop/I have been invited to present at the 2nd AUEB Sports Analytics Workshop in Athens, Greece.GSE Graduate course: 'Generalized linear models with massive data'
http://www.ikosmidis.com/post/news-bigglm_july17/
Sat, 15 Jul 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-bigglm_july17/On 18 and 19 October 2017, I will be giving a course on statistical learning of generalized linear models with massive amounts of data at Barcelona GSE Data Science.Improved estmation with ordinal data at CEN ISBS 2017 Joint Conference
http://www.ikosmidis.com/post/news-cenisbs2017/
Fri, 02 Jun 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-cenisbs2017/I have been invited to present my work on Improved estimation in cumulative link models at the CEN ISBS 2017 Joint Conference in Vienna, Austria. The talk is on 31 August 2017, 09:30 - 10:00.
Update (31 August 2017): The slides for the talk are available here.My brglm2 R package is on CRAN
http://www.ikosmidis.com/post/news-brglm2/
Thu, 25 May 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-brglm2/brglm2 provides various methods for mean and median bias reduction in the estimation of generalized linear models, along with pre-fit and post-fit methods for the detection of separation and of infinite maximum likelihood estimates in binomial response generalized linear models. See the package vignettes and the brglm2 GitHub page for details.My enrichwith R package is on CRAN
http://www.ikosmidis.com/post/news-enrichwith/
Thu, 18 May 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-enrichwith/enrichwith provides the “enrich” method (verb) to enrich list-like R objects with new, relevant components. The current version can enrich objects of class ‘family’, ‘link-glm’, ‘lm’ and ‘glm’. The package also provides the ‘enriched_glm’ function that results in objects that carry various useful methods for generalized linear models (simulate, observed and expected information matrix, and model densities, probabilities, and quantiles at arbitrary parameter values). See the package vignettes and the enrichwith GitHub page for more details.Meta-regression paper appears in Biometrika
http://www.ikosmidis.com/post/news-meta-analysis-paper/
Tue, 28 Mar 2017 00:00:00 +0000http://www.ikosmidis.com/post/news-meta-analysis-paper/Improving the accuracy of likelihood-based inference in meta-analysis and meta-regression (joint work with Annamaria Guolo and Cristiano Varin) appeared online in Biometrika.GLMs with massive data in Barcelona and at Innsbruck
http://www.ikosmidis.com/post/news-bigglm_november16/
Sat, 12 Nov 2016 00:00:00 +0000http://www.ikosmidis.com/post/news-bigglm_november16/I will be giving a course on statistical learning of generalized linear models with massive amounts of data at the Barcelona GSE Data Science (November 2016) and at the Faculty of Economics and Statistics, University of Innsbruck (January 2017).Inaugural Turing Fellow
http://www.ikosmidis.com/post/news-turing_appointment/
Wed, 05 Oct 2016 00:00:00 +0000http://www.ikosmidis.com/post/news-turing_appointment/I have been selected to be one of the inaugural Turing Fellows at the newly founded Alan Turing Institute.3rd UCL Workshop on the theory of big data
http://www.ikosmidis.com/post/news-3rd-big-data-workshop/
Tue, 06 Sep 2016 00:00:00 +0000http://www.ikosmidis.com/post/news-3rd-big-data-workshop/I am co-organising the 3rd UCL Workshop on the Theory of Big Data (26 - 28 June 2017, London, UK).Bias in Parametric Estimation in Top 10 of WIREs Computational Statistics
http://www.ikosmidis.com/post/news-bias-estimation/
Sun, 22 May 2016 00:00:00 +0000http://www.ikosmidis.com/post/news-bias-estimation/My article Bias in parametric estimation: reduction and useful side-effects features in top ten most accessed articles of WIREs Computational Statistics in 2015.Turing scoping workshop: Data Science for data-rich sports
http://www.ikosmidis.com/post/news-turing-scoping-workshop/
Fri, 11 Sep 2015 00:00:00 +0000http://www.ikosmidis.com/post/news-turing-scoping-workshop/The Alan Turing Institute funded the scoping workshop Data Science for data-rich sports, which I co-proposed (together with David Firth and Anjali Mazumder at Warwick, Richard Samworth at Cambridge, and Kostas Zygalakis at Edinburgh). The workshop will take place on 2 and 3 November 2015. Details can be found here and here.Model-based clustering using copulas is in Statistics and Computing
http://www.ikosmidis.com/post/news-copulas-paper/
Wed, 28 Jan 2015 00:00:00 +0000http://www.ikosmidis.com/post/news-copulas-paper/Model-based clustering using copulas with applications appeared online in Statistics and Computing.GLMs at the 2nd R Spring Court @ AUEB
http://www.ikosmidis.com/post/news-glms-aueb2/
Sat, 03 Jan 2015 00:00:00 +0000http://www.ikosmidis.com/post/news-glms-aueb2/On March 2015, I will be giving a course at the 2nd R Spring Course @ AUEB at Athens, Greece.2nd UCL Workshop on the theory of big data
http://www.ikosmidis.com/post/news-2nd-big-data-workshop/
Mon, 06 Oct 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-2nd-big-data-workshop/I am co-organising the 2nd UCL Workshop on the Theory of Big Data, which will take place in January 2016 in London, UK.Research contract with English Institute of Sport
http://www.ikosmidis.com/post/news-eis-project/
Sun, 05 Oct 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-eis-project/I will be leading a collaborative project with English Institute of Sport on developing statistical methods for sport performance data. The collaboration includes two year’s funding for the appointment of a Research Associate in Statistics. News item here.MAPS Faculty Teaching Award
http://www.ikosmidis.com/post/news-maps-teaching-award/
Fri, 29 Aug 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-maps-teaching-award/I was awarded the MAPS Faculty Teaching Award in the academic staff category. News items here and here.Bias in Parametric Estimation paper appears in WIREs Computational Statistics
http://www.ikosmidis.com/post/news-bias-estimation-appears/
Thu, 10 Apr 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-bias-estimation-appears/Bias in parametric estimation: reduction and useful side-effects appeared in WIRE Computational Statistics.GLMs at the 1st R Summaer School @ AUEB
http://www.ikosmidis.com/post/news-glms-aueb1/
Fri, 07 Mar 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-glms-aueb1/On 23-27 June 2015, I will be giving a course on generalized linear models at the 1st R Summer School @ AUEB in Athens, Greece.Improved estimation in cumulative link models in RSSB
http://www.ikosmidis.com/post/news-imroved-clm-appears/
Thu, 27 Feb 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-imroved-clm-appears/Improved estimation in cumulative link models appeared in the Journal of the Royal Statistical Society: Series B.Handling nuisance parameters at ERCIM 2014
http://www.ikosmidis.com/post/news-handling-nuisance-parameters/
Wed, 12 Feb 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-handling-nuisance-parameters/I am organising the session “Handling nuisance parameters: Advances towards optimal inference in statistics and econometrics” at ERCIM 2014 (6th - 8th December 2014, Pisa, Italy).1st UCL Workshop on the theory of big data
http://www.ikosmidis.com/post/news-1st-big-data-workshop/
Mon, 06 Jan 2014 00:00:00 +0000http://www.ikosmidis.com/post/news-1st-big-data-workshop/I am co-organising the 1st UCL Workshop on the Theory of Big Data (7th - 9th January 2015, London, UK).