Podcasts about bayesian methods

Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Matt emphasizes the importance of Bayesian statistics in scenarios with limited data.Communicating insights to coaches is a crucial skill for data analysts.Building a data team requires understanding the needs of the coaching staff.Player recruitment is a significant focus in football analytics.The integration of data science in sports is still evolving.Effective data modeling must consider the practical application in games.Collaboration between data analysts and coaches enhances decision-making.Having a robust data infrastructure is essential for efficient analysis.The landscape of sports analytics is becoming increasingly competitive. Player recruitment involves analyzing various data models.Biases in traditional football statistics can skew player evaluations.Statistical techniques should leverage the structure of football data.Tracking data opens new avenues for understanding player movements.The role of data analysis in football will continue to grow.Aspiring analysts should focus on curiosity and practical experience.Chapters:00:00 Introduction to Football Analytics and Matt's Journey04:54 The Role of Bayesian Methods in Football10:20 Challenges in Communicating Data Insights17:03 Building Relationships with Coaches22:09 The Structure of the Data Team at Como26:18 Focus on Player Recruitment and Transfer Strategies28:48 January Transfer Window Insights30:54 Biases in Football Data Analysis34:11 Comparative Analysis of Men's and Women's Football36:55 Statistical Techniques in Football Analysis42:48 The Impact of Tracking Data on Football Analysis45:49 The Future of Data-Driven Football Strategies47:27 Advice for Aspiring Football Analysts

Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!My Intuitive Bayes Online Courses1:1 Mentorship with meOur theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Bob's research focuses on corruption and political economy.Measuring corruption is challenging due to the unobservable nature of the behavior.The challenge of studying corruption lies in obtaining honest data.Innovative survey techniques, like randomized response, can help gather sensitive data.Non-traditional backgrounds can enhance statistical research perspectives.Bayesian methods are particularly useful for estimating latent variables.Bayesian methods shine in situations with prior information.Expert surveys can help estimate uncertain outcomes effectively.Bob's novel, 'The Bayesian Heatman,' explores academia through a fictional lens.Writing fiction can enhance academic writing skills and creativity.The importance of community in statistics is emphasized, especially in the Stan community.Real-time online surveys could revolutionize data collection in social science.Chapters:00:00 Introduction to Bayesian Statistics and Bob Kubinec06:01 Bob's Academic Journey and Research Focus12:40 Measuring Corruption: Challenges and Methods18:54 Transition from Government to Academia26:41 The Influence of Non-Traditional Backgrounds in Statistics34:51 Bayesian Methods in Political Science Research42:08 Bayesian Methods in COVID Measurement51:12 The Journey of Writing a Novel01:00:24 The Intersection of Fiction and AcademiaThank you to my Patrons for making this episode possible!Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, Tim Gasser, James Wade, Tradd Salvo, William Benton, James Ahloy, Robin Taylor,, Chad Scherrer, Zwelithini Tunyiswa, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian Costley, Ally Salim, Larry Gill, Ian Moran, Paul Oreto, Colin Caprani, Colin Carroll, Nathaniel Burbank, Michael Osthege, Rémi Louf, Clive Edelsten, Henri Wallen, Hugo Botha, Vinh Nguyen, Marcin Elantkowski, Adam C. Smith, Will Kurt, Andrew Moskowitz, Hector Munoz, Marco Gorelli, Simon Kessell,...

Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!My Intuitive Bayes Online Courses1:1 Mentorship with meOur theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Use mini-batch methods to efficiently process large datasets within Bayesian frameworks in enterprise AI applications.Apply approximate inference techniques, like stochastic gradient MCMC and Laplace approximation, to optimize Bayesian analysis in practical settings.Explore thermodynamic computing to significantly speed up Bayesian computations, enhancing model efficiency and scalability.Leverage the Posteriors python package for flexible and integrated Bayesian analysis in modern machine learning workflows.Overcome challenges in Bayesian inference by simplifying complex concepts for non-expert audiences, ensuring the practical application of statistical models.Address the intricacies of model assumptions and communicate effectively to non-technical stakeholders to enhance decision-making processes.Chapters:00:00 Introduction to Large-Scale Machine Learning11:26 Scalable and Flexible Bayesian Inference with Posteriors25:56 The Role of Temperature in Bayesian Models32:30 Stochastic Gradient MCMC for Large Datasets36:12 Introducing Posteriors: Bayesian Inference in Machine Learning41:22 Uncertainty Quantification and Improved Predictions52:05 Supporting New Algorithms and Arbitrary Likelihoods59:16 Thermodynamic Computing01:06:22 Decoupling Model Specification, Data Generation, and InferenceThank you to my Patrons for making this episode possible!Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, Tim Gasser, James Wade, Tradd Salvo, William Benton, James Ahloy, Robin Taylor,, Chad Scherrer, Zwelithini Tunyiswa, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian Costley, Ally Salim, Larry Gill, Ian Moran, Paul Oreto, Colin Caprani, Colin Carroll, Nathaniel Burbank, Michael Osthege, Rémi Louf, Clive Edelsten, Henri Wallen, Hugo Botha, Vinh Nguyen, Marcin Elantkowski, Adam C. Smith, Will Kurt, Andrew Moskowitz, Hector Munoz, Marco Gorelli, Simon Kessell, Bradley Rode, Patrick Kelley, Rick Anderson, Casper de Bruin, Philippe Labonde, Michael Hankin, Cameron Smith, Tomáš Frýda, Ryan Wesslen, Andreas Netti, Riley King, Yoshiyuki Hamajima, Sven De Maeyer, Michael DeCrescenzo, Fergal

Bayesian methods take the spotlight in this episode with Alex Andorra, co-founder of PyMC Labs, and Jon Krohn. Learn how Bayesian techniques handle tough problems, make the most of prior knowledge, and work wonders with limited data. Alex and Jon break down essentials like PyMC, PyStan, and NumPyro libraries, show how to boost model efficiency with PyTensor, and talk about using ArviZ for top-notch diagnostics and visualizations. Plus, get into advanced modeling with Gaussian Processes. This episode is brought to you by Crawlbase (https://crawlbase.com), the ultimate data crawling platform. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • Practical introduction to Bayesian statistics [04:54] • Definition and significance of epistemology [17:52] • Explanation of PyMC and Monte Carlo methods [27:57] • How to get started with Bayesian modeling and PyMC [34:26] • PyMC Labs and its consulting services [50:50] • ArviZ for post-modeling diagnostics and visualization [01:02:23] • Gaussian processes and their applications [01:09:02] Additional materials: www.superdatascience.com/793

In this episode we hear from David Denison, Deputy CIO of Hedge Fund Florin Court Capital, a diversified systematic asset manager.  David discusses the philosophy behind trading multiple diversified asset classes across the power / electricity, commodity, interest rate swap and FX markets, the dynamics of a market with fewer speculating players and more producer players, the periods in which trend-following has been particularly successful historically, the benefits of trend-following in mitigating behavioral biases such as anchoring, and the growth in alternative CTAs as a HF asset class.  David Denison is in discussion with Eloise Goulder, Head of the Data Assets & Alpha Group. Shownotes: To learn more about Florin Court Capital, see https://florincourt.com/ To learn more about David's book, see: “Bayesian Methods for Nonlinear Classification and Regression”, Wiley, 2002. This episode was recorded on October 16th, 2023. The views expressed in this podcast may not necessarily reflect the views of J.P. Morgan Chase & Co and its affiliates (together “J.P. Morgan”), they are not the product of J.P. Morgan's Research Department and do not constitute a recommendation, advice, or an offer or a solicitation to buy or sell any security or financial instrument.  This podcast is intended for institutional and professional investors only and is not intended for retail investor use, it is provided for information purposes only. Referenced products and services in this podcast may not be suitable for you and may not be available in all jurisdictions.  J.P. Morgan may make markets and trade as principal in securities and other asset classes and financial products that may have been discussed.  For additional disclaimers and regulatory disclosures, please visit: www.jpmorgan.com/disclosures/salesandtradingdisclaimer. For the avoidance of doubt, opinions expressed by any external speakers are the personal views of those speakers and do not represent the views of J.P. Morgan.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Are Bayesian methods guaranteed to overfit?, published by Ege Erdil on June 17, 2023 on LessWrong. Yuling Yao argues that Bayesian models are guaranteed to overfit. He summarizes his point as follows: I have a different view. Bayesian model does overfit. Moreover, Bayes is guaranteed to overfit, regardless of the model (being correct or wrong) or the prior ( “strong” or uninformative). Moreover, Bayes is guaranteed to overfit on every realization of training data, not just in expectation. Moreover, Bayes is guaranteed to overfit on every single point of the training data, not just in the summation. He uses the following definition of "overfitting": a model "overfits" some data if its out-of-sample log loss exceeds its within-sample log loss. Interpreted in a different way, this is equivalent to saying that the model assigns higher probability to a data point after updating on it than before. Andrew Gelman makes the point that any proper fitting procedure whatsoever has this property, and alternative methods "overfit" more than ideal Bayesian methods. I think the proper way to interpret the results is not that Bayesian methods are guaranteed to overfit but that the definition of "overfitting" used by Yuling Yao, while intuitively plausible at first glance, is actually poor. Still, proving the fact that Bayesian methods indeed must "overfit" in his sense is an interesting exercise. I tried understanding his derivation of this and gave up - I present an original derivation of the same fact below that I hope is clearer. Derivation Suppose we have a model parametrized by parameters θ and the probability of seeing some data y according to our model is P(y|θ). Now, suppose we draw n independent samples y1,y2,.,yn. Denote this whole data vector by y, and denote the data vector with the ith sample omitted by y−i. Under Bayesian inference, the within-sample probability of observing the value yi in the next sample we draw is P(yn+1=yi|y)=∫θP(θ|y)P(yi|θ)dθ On the other hand, Bayes says that P(θ|y)=P(θ|y−i,yi)=P(θ|y−i)P(yi|y−i,θ)P(yi|y−i)=P(θ|y−i)P(yi|θ)P(yi|y−i) Plugging in gives P(yn+1=yi|y)=∫θP(θ|y−i)P(yi|θ)2P(yi|y−i)dθ or P(yn+1=yi|y)P(yi|y−i)=Eθ∼P(θ|y−i)[P(yi|θ)2] We can decompose the expectation of the squared probability on the right hand side using the definition of variance as follows: P(yn+1=yi|y)P(yi|y−i)=Eθ∼P(θ|y−i)[P(yi|θ)]2+varθ∼P(θ|y−i)(P(yi|θ))=P(yi|y−i)2+varθ∼P(θ|y−i)(P(yi|θ)) where I've used the fact that Eθ∼P(θ|y−i)[P(yi|θ)]=∫θP(θ|y−i)P(yi|θ)dθ=P(yi|y−i) to get rid of the expectation. The variance term on the right hand side is nonnegative by definition as it's a variance, and it's strictly positive as long as there's any uncertainty in our beliefs about θ after seeing the data y−i that would influence our probability estimate of observing yi next. This will be the case in almost all nondegenerate situations, and if so, we obtain the strict inequality P(yn+1=yi|y)>P(yi|y−i) What does this mean? The finding is intuitively obvious, but poses some challenges to formally defining the notion of overfitting. This is essentially because the ideal amount of fitting for a model to do on some data is nonzero, and overfitting should be "exceeding" this level of ideal fitting. In practice, though, it's difficult to know what is the "appropriate" amount of fitting for a model to be doing. Bayesian inference is ideal if the true model is within the class of models under consideration, but it might fail in unexpected ways if it's not, which is almost always the case in practice. I think the lesson to draw from this is that overfitting is a relative concept and claiming that a particular method "overfits" the data doesn't make too much sense without a point of reference in mind. If people have alternative ways of trying to construct an absolute notion of overfitting with the a...

Link to original articleWelcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Are Bayesian methods guaranteed to overfit?, published by Ege Erdil on June 17, 2023 on LessWrong. Yuling Yao argues that Bayesian models are guaranteed to overfit. He summarizes his point as follows: I have a different view. Bayesian model does overfit. Moreover, Bayes is guaranteed to overfit, regardless of the model (being correct or wrong) or the prior ( “strong” or uninformative). Moreover, Bayes is guaranteed to overfit on every realization of training data, not just in expectation. Moreover, Bayes is guaranteed to overfit on every single point of the training data, not just in the summation. He uses the following definition of "overfitting": a model "overfits" some data if its out-of-sample log loss exceeds its within-sample log loss. Interpreted in a different way, this is equivalent to saying that the model assigns higher probability to a data point after updating on it than before. Andrew Gelman makes the point that any proper fitting procedure whatsoever has this property, and alternative methods "overfit" more than ideal Bayesian methods. I think the proper way to interpret the results is not that Bayesian methods are guaranteed to overfit but that the definition of "overfitting" used by Yuling Yao, while intuitively plausible at first glance, is actually poor. Still, proving the fact that Bayesian methods indeed must "overfit" in his sense is an interesting exercise. I tried understanding his derivation of this and gave up - I present an original derivation of the same fact below that I hope is clearer. Derivation Suppose we have a model parametrized by parameters θ and the probability of seeing some data y according to our model is P(y|θ). Now, suppose we draw n independent samples y1,y2,.,yn. Denote this whole data vector by y, and denote the data vector with the ith sample omitted by y−i. Under Bayesian inference, the within-sample probability of observing the value yi in the next sample we draw is P(yn+1=yi|y)=∫θP(θ|y)P(yi|θ)dθ On the other hand, Bayes says that P(θ|y)=P(θ|y−i,yi)=P(θ|y−i)P(yi|y−i,θ)P(yi|y−i)=P(θ|y−i)P(yi|θ)P(yi|y−i) Plugging in gives P(yn+1=yi|y)=∫θP(θ|y−i)P(yi|θ)2P(yi|y−i)dθ or P(yn+1=yi|y)P(yi|y−i)=Eθ∼P(θ|y−i)[P(yi|θ)2] We can decompose the expectation of the squared probability on the right hand side using the definition of variance as follows: P(yn+1=yi|y)P(yi|y−i)=Eθ∼P(θ|y−i)[P(yi|θ)]2+varθ∼P(θ|y−i)(P(yi|θ))=P(yi|y−i)2+varθ∼P(θ|y−i)(P(yi|θ)) where I've used the fact that Eθ∼P(θ|y−i)[P(yi|θ)]=∫θP(θ|y−i)P(yi|θ)dθ=P(yi|y−i) to get rid of the expectation. The variance term on the right hand side is nonnegative by definition as it's a variance, and it's strictly positive as long as there's any uncertainty in our beliefs about θ after seeing the data y−i that would influence our probability estimate of observing yi next. This will be the case in almost all nondegenerate situations, and if so, we obtain the strict inequality P(yn+1=yi|y)>P(yi|y−i) What does this mean? The finding is intuitively obvious, but poses some challenges to formally defining the notion of overfitting. This is essentially because the ideal amount of fitting for a model to do on some data is nonzero, and overfitting should be "exceeding" this level of ideal fitting. In practice, though, it's difficult to know what is the "appropriate" amount of fitting for a model to be doing. Bayesian inference is ideal if the true model is within the class of models under consideration, but it might fail in unexpected ways if it's not, which is almost always the case in practice. I think the lesson to draw from this is that overfitting is a relative concept and claiming that a particular method "overfits" the data doesn't make too much sense without a point of reference in mind. If people have alternative ways of trying to construct an absolute notion of overfitting with the a...

Brad Carlin is a statistical researcher, methodologist, consultant, and instructor.  He currently serves as Senior Advisor for Data Science and Statistics at PharmaLex, an international pharmaceutical consulting firm.  Prior to this, he spent 27 years on the faculty of the Division of Biostatistics at the University of Minnesota School of Public Health, serving as division head for 7 of those years.  He has also held visiting positions at Carnegie Mellon University, Medical Research Council Biostatistics Unit, Cambridge University (UK), Medtronic Corporation, HealthPartners Research Foundation, the M.D Anderson Cancer Center, and AbbVie Pharmaceuticals.   He has published more than 185 papers in refereed books and journals, and has co-authored three popular textbooks: “Bayesian Methods for Data Analysis” with Tom Louis, “Hierarchical Modeling and Analysis for Spatial Data” with Sudipto Banerjee and Alan Gelfand, and "Bayesian Adaptive Methods for Clinical Trials" with Scott Berry, J. Jack Lee, and Peter Muller.  From 2006-2009 he served as editor-in-chief of Bayesian Analysis, the official journal of the International Society for Bayesian Analysis (ISBA).  During his academic career, he served as primary dissertation adviser for 20 PhD students.  Dr. Carlin has extensive experience teaching short courses and tutorials, and won both teaching and mentoring awards from the University of Minnesota. During his spare time, Brad is a health musician and bandleader, providing keyboards, guitar, and vocals in a variety of venues.

The big problems with classic hypothesis testing are well-known. And yet, a huge majority of statistical analyses are still conducted this way. Why is it? Why are things so hard to change? Can you even do (and should you do) hypothesis testing in the Bayesian framework? I guess if you wanted to name this episode in a very Marvelian way, it would be “Bayes factors against the p-values of madness” — but we won't do that, it wouldn't be appropriate, would it? Anyways, in this episode, I'll talk about all these very light and consensual topics with Eric-Jan Wagenmakers, a professor at the Psychological Methods Unit of the University of Amsterdam. For almost two decades, EJ has staunchly advocated the use of Bayesian inference in psychology. In order to lower the bar for the adoption of Bayesian methods, he is coordinating the development of JASP, an open-source software program that allows practitioners to conduct state-of-the-art Bayesian analyses with their mouse — the one from the computer, not the one from Disney. EJ has also written a children's book on Bayesian inference with the title “Bayesian thinking for toddlers”. Rumor has it that he is also working on a multi-volume series for adults — but shhh, that's a secret! EJ's lab publishes regularly on a host of Bayesian topics, so check out his website, particularly when you are interested in Bayesian hypothesis testing. The same goes for his blog by the way, “BayesianSpectacles”. Wait, what's that? EJ is telling me that he plays chess, squash, and that, most importantly, he enjoys watching arm wrestling videos on YouTube — yet another proof that, yes, you can find everything on YouTube. Our theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work at https://bababrinkman.com/ (https://bababrinkman.com/) ! Thank you to my Patrons for making this episode possible! Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, Tim Gasser, James Wade, Tradd Salvo, Adam Bartonicek, William Benton, Alan O'Donnell, Mark Ormsby, James Ahloy, Robin Taylor, Thomas Wiecki, Chad Scherrer, Nathaniel Neitzke, Zwelithini Tunyiswa, Elea McDonnell Feit, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian Costley, Ally Salim, Larry Gill, Joshua Duncan, Ian Moran, Paul Oreto, Colin Caprani, George Ho, Colin Carroll, Nathaniel Burbank, Michael Osthege, Rémi Louf, Clive Edelsten, Henri Wallen, Hugo Botha, Vinh Nguyen, Raul Maldonado, Marcin Elantkowski, Adam C. Smith, Will Kurt, Andrew Moskowitz, Hector Munoz, Marco Gorelli, Simon Kessell, Bradley Rode, Patrick Kelley, Rick Anderson, Casper de Bruin, Philippe Labonde, Matthew McAnear, Michael Hankin, Cameron Smith, Luis Iberico, Tomáš Frýda, Ryan Wesslen, Andreas Netti, Riley King, Aaron Jones, Yoshiyuki Hamajima, Sven De Maeyer, Michael DeCrescenzo, Fergal M, Mason Yahr, Naoya Kanai, Steven Rowland and Aubrey Clayton. Visit https://www.patreon.com/learnbayesstats (https://www.patreon.com/learnbayesstats) to unlock exclusive Bayesian swag ;) Links from the show: EJ's website: http://ejwagenmakers.com/ (http://ejwagenmakers.com/) EJ on Twitter: https://twitter.com/EJWagenmakers (https://twitter.com/EJWagenmakers) “Bayesian Cognitive Modeling” book website: https://bayesmodels.com/ (https://bayesmodels.com/) Port of “Bayesian Cognitive Modeling” to PyMC: https://github.com/pymc-devs/pymc-resources/tree/main/BCM (https://github.com/pymc-devs/pymc-resources/tree/main/BCM) EJ's blog: http://www.bayesianspectacles.org/ (http://www.bayesianspectacles.org/) JASP software website: https://jasp-stats.org/ (https://jasp-stats.org/) Bayesian Thinking for Toddlers: https://psyarxiv.com/w5vbp/ (https://psyarxiv.com/w5vbp/) LBS #31, Bayesian Cognitive Modeling & Decision-Making with Michael Lee: https://www.learnbayesstats.com/episode/31-bayesian-cognitive-modeling-michael-lee...

On this episode of On the Evidence, Mathematica's Mariel Finucane and John Deke join Tim Day of the Center for Medicare & Medicaid Innovation to discuss the application of evidence-informed Bayesian methods that not only confirm whether a policy or program works, but for whom. Learn more about Mathematica's work using evidence-based Bayesian methods in applied policy research: https://mathematica.org/features/bayesian-methods Read a brief about using a Bayesian framework for interpreting findings from impact evaluations prepared by Mariel Finucane and John Deke for the Office of Planning, Research and Evaluation at the Administration for Children and Families: mathematica.org/publications/moving-beyond-statistical-significance-the-basie-bayesian-interpretation-of-estimates-framework Read a paper co-authored by Mariel Finucane that compares Bayesian methods with the traditional frequentist approach to estimate the effects of a Centers for Medicare & Medicaid Services demonstration on Medicare spending: mathematica.org/publications/revolutionizing-estimation-and-inference-for-program-evaluation-using-bayesian-methods Read a paper co-authored by Tim Day describing an experiment to provide evidence that would be useful to policymakers and other decision makers through an interactive data visualization dashboard, presenting results from both frequentist and Bayesian analyses: https://www.researchgate.net/publication/335169870_Making_Evidence_Actionable_Interactive_Dashboards_Bayes_and_Health_Care_Innovation Read Emily Oster's newsletter article about why and how she applies Bayes's Rule to interpret new evidence in the context of existing evidence, including a recent study (https://emilyoster.substack.com/p/does-pre-k-really-hurt-future-test) about the effects of a preschool program in Tennessee on future student test scores: https://emilyoster.substack.com/p/bayes-rule-is-my-faves-rule

In this episode we interview Dr. Ghassan Hamra and talk about all things Bayesian. If you’re like us, you have likely been trained in traditional, frequentist approaches to statistics and have always wondered what the big deal is about Bayesian approaches. Well, have no fear, Dr. Hamra is here to explain it all. In this episode we cover a range of topics introducing Bayesian analyses, including how Bayesian and frequentist statistics differ, the concept of integrating a prior into your analyses, and whether Bayesian statistics are really a “subjective” approach (**spoiler alert: they’re not). After listening to this podcast, if you’re interested in learning more about Bayesian analyses some links are included below: MacLehose, R.F., Hamra, G.B. Applications of Bayesian Methods to Epidemiologic Research. Curr Epidemiol Rep 1, 103–109 (2014). https://doi.org/10.1007/s40471-014-0019-z Hamra GB, MacLehose RF, Cole SR. Sensitivity analyses for sparse-data problems-using weakly informative bayesian priors. Epidemiology. 2013;24(2):233-239. doi:10.1097/EDE.0b013e318280db1d Website with links to Dr. Hamra’s publications and presentations/tutorials: http://ghassanbhamra-phd.org/publications http://ghassanbhamra-phd.org/presentations-and-such Series of articles by Sander Greenland on Bayesian methods for epidemiology: Sander Greenland, Bayesian perspectives for epidemiological research: I. Foundations and basic methods, International Journal of Epidemiology, Volume 35, Issue 3, June 2006, Pages 765–775, https://doi.org/10.1093/ije/dyi312 Sander Greenland, Bayesian perspectives for epidemiological research. II. Regression analysis, International Journal of Epidemiology, Volume 36, Issue 1, February 2007, Pages 195–202, https://doi.org/10.1093/ije/dyl289 Sander Greenland, Bayesian perspectives for epidemiologic research: III. Bias analysis via missing-data methods, International Journal of Epidemiology, Volume 38, Issue 6, December 2009, Pages 1662–1673, https://doi.org/10.1093/ije/dyp278 MacLehose RF, Gustafson P. Is probabilistic bias analysis approximately Bayesian?. Epidemiology. 2012;23(1):151-158. doi:10.1097/EDE.0b013e31823b539c

Part 2 of a three part episode with Martin Ho and Greg Maislin, talking about the ASA Section on Medical Devices and Diagnostics (MDD).  This part discusses Bayesian Methods and Digital Health Initiatives. The other two parts of this episodes cover: Part 1: MDD Section Activities Part 3: The MDD Idea Exchange and Bayesian p-values

Part 1 of a three part episode with Martin Ho and Greg Maislin, talking about the ASA Section on Medical Devices and Diagnostics (MDD).  This part discusses the MDD Section Activities.    The other two parts of this episodes will cover: Part 2: Bayesian Methods and Digital Health Initiatives Part 3: The MDD Idea Exchange and Bayesian p-values

Part 3 of a three part episode with Martin Ho and Greg Maislin, talking about the ASA Section on Medical Devices and Diagnostics (MDD).  This part discusses the MDD Idea Exchange and Bayesian p-values. The other two parts of this episodes cover: Part 1: MDD Section Activities Part 2: Bayesian Methods and Digital Health Initiatives

This week we open with a critical take on our current system of disseminating scientific research, specifically focusing on the prevalence of -- and dependence on -- medical writers. In the second half of the episode, we interview Dr. Allen Pannell of the Haslam College of Business at the University of Tennessee on using a Frequentist approach vs a Bayesian approach in the context of a single clinical trial. Frequentist vs Bayesian: doi.org/10.1136/bmjopen-2018-024256 Dr. Pannell's breast cancer support group: https://www.breastconnect.org/ Dr. Pannell's research project: http://tinyurl.com/peerErPrHer2 Back us on Patreon! www.patreon.com/plenarysession

Professor Maurice Pagnucco is the Deputy Dean of Education, and the Head of the School of Computer Science and Engineering at the University of New South Wales (UNSW) in Sydney Australia. A talk which turned into a Brief History of Machine Learning & using Prof Pagnucco's wide & deep knowledge of the field - I asked him a couple of ML questions covering topics including Expert Systems, Bayesian Methods, Causal Inference, the Millennium Prize and the future of ML.

Branden Fitelson (Rutgers University) gives a talk at the MCMP Workshop on Bayesian Methods in Philosophy titled "Accuracy & Coherence". Abstract: In this talk, I will explore a new way of thinking about the relationship between accuracy norms and coherence norms in epistemology (generally). In the first part of the talk, I will apply the basic ideas to qualitative judgments (belief and disbelief). This will lead to an interesting coherence norm for qualitative judgments (but one which is weaker than classical deductive consistency). In the second part of the talk, I will explain how the approach can be applied to comparative confidence judgments. Again, this will lead to coherence norms that are weaker than classical (comparative probabilistic) coherence norms. Along the way, I will explain how evidential norms can come into conflict with even the weaker coherence norms suggested by our approach.

Vincenzo Crupi (MCMP/LMU) gives a talk at the MCMP Workshop on Bayesian Methods in Philosophy titled "Formal epistemological explication (news for the Bayesian agenda)".

Niki Pfeifer (MCMP/LMU) gives a talk at the MCMP Workshop on Bayesian Methods in Philosophy titled "Applying coherence based probability logic to philosophical problems".

Hannes Leitgeb (MCMP/LMU) gives a talk at the MCMP Workshop on Bayesian Methods in Philosophy titled "The Lockean Thesis Revisited".

Charles B. Cross (University of Georgia) gives a talk at the MCMP Workshop on Bayesian Methods in Philosophy titled "Knowledge about Probability in the Monty Hall Problem".

Today's guest is Cameron Davidson-Pilon. Cameron has a masters degree in quantitative finance from the University of Waterloo. Think of it as statistics on stock markets. For the last two years he's been the team lead of data science at Shopify. He's the founder of dataoragami.net which produces screencasts teaching methods and techniques of applied data science. He's also the author of the just released in print book Bayesian Methods for Hackers: Probabilistic Programming and Bayesian Inference, which you can also get in a digital form. This episode focuses on the topic of Bayesian A/B Testing which spans just one chapter of the book. Related to today's discussion is the Data Origami post The class imbalance problem in A/B testing. Lastly, Data Skeptic will be giving away a copy of the print version of the book to one lucky listener who has a US based delivery address. To participate, you'll need to write a review of any site, book, course, or podcast of your choice on datasciguide.com. After it goes live, tweet a link to it with the hashtag #WinDSBook to be given an entry in the contest. This contest will end November 20th, 2015, at which time I'll draw a single randomized winner and contact them for delivery details via direct message on Twitter.

Karen Price talks about all things Bayesian in the world of pharmaceutical development: methodology, the DIA Bayesian Scientific Working Group, and the special Bayesian issue of Pharmaceutical Statistics.

An outline for questions I hope to answer: What is Bayes’ Rule? (lecture portion) ► What is the likelihood? ► What is the prior distribution? ► How should I choose it? ► Why use a conjugate prior? ► What is a subjective versus objective prior? ► What is the posterior distribution? ► How do I use it to make statistical inference? ► How is this inference different from frequentist/classical inference? ► What computational tools do I need in order to make inference? How can I use R to do regression in a Bayesian paradigm? (computer portion) ► What libraries in R support Bayesian analysis? ► How do I use some of these libraries? ► How do I interpret the output? ► How do I produce diagnostic plots? ► What common topics do these libraries not support? ► How can I do them myself? ► How can LISA help me? ► What resources are available to help me Bayesian methods in R? Course files available here: www.lisa.stat.vt.edu/?q=node/3382.

We’ll discuss some basic concepts and vocabulary in Bayesian statistics such as the likelihood, prior and posterior distributions, and how they relate to Bayes’ Rule. R statistical software will be used to discuss how parameter estimation and inference changes in a Bayesian paradigm versus in a classical paradigm, with a particular focus on applications using regression. Course files available here: www.lisa.stat.vt.edu/?q=node/1784.

The cosmic large scale structure is of special relevance for testing current cosmological theories about the origin and evolution of the Universe. Throughout cosmic history, it evolved from tiny quantum fluctuations, generated during the early epoch of inflation, to the filamentary cosmic web presently observed by our telescopes. Observations and analyses of this large scale structure will hence test this picture, and will provide valuable information on the processes of cosmic structure formation as well as they will reveal the cosmological parameters governing the dynamics of the Universe. Beside measurements of the cosmic microwave backround, galaxy observations are of particular interest to modern precision cosmology. They are complementary to many other sources of information, such as cosmic microwave background experiments, since they probe a different epoch. Galaxies report the cosmic evolution over an enormous period ranging from the end of the epoch of reionization, when luminous objects first appeared, till today. For this reason, galaxy surveys are excellent probes of the dynamics and evolution of the Universe. Especially the Sloan Digital Sky Survey is one of the most ambitious surveys in the history of astronomy. It provides measurements of 930,000 galaxy spectra as well as the according angular and redshift positions of galaxies over an area which covers more than a quarter of the sky. This enormous amount of precise data allows for an unprecedented access to the three dimensional cosmic matter distribution and its evolution. However, observables, such as positions and properties of galaxies, provide only an inaccurate picture of the cosmic large scale structure due to a variety of statistical and systematic observational uncertainties. In particular, the continuous cosmic density field is only traced by a set of discrete galaxies introducing statistical uncertainties in the form of Poisson distributed noise. Further, galaxy surveys are subject to a variety of complications such as instrumental limitations or the nature of the observation itself. The solution to the underlying problem of characterizing the large scale structure in the Universe therefore requires a statistical approach. The main theme of this PhD-thesis is the development of new Bayesian data analysis methods which provide a complete statistical characterization and a detailed cosmographic description of the large scale structure in our Universe. The required epistemological concepts, the mathematical framework of Bayesian statistics as well as numerical considerations are thoroughly discussed. On this basis two Bayesian data analysis computer algorithms are developed. The first of which is called ARES (Algorithm for REconstruction and Sampling). It aims at the joint inference of the three dimensional density field and its power-spectrum from galaxy observations. The ARES algorithm accurately treats many observational systematics and statistical uncertainties, such as the survey geometry, galaxy selection effects, blurring effects and noise. Further, ARES provides a full statistical characterization of the three dimensional density field, the power-spectrum and their joint uncertainties by exploring the high dimensional space of their joint posterior via a very efficient Gibbs sampling scheme. The posterior is the probability of the model given the observations and all other available informations. As a result, ARES provides a sampled representation of the joint posterior, which conclusively characterizes many of the statistical properties of the large scale structure. This probability distribution allows for a variety of scientific applications, such as reporting any desired statistical summary or testing of cosmological models via Bayesian model comparison or Bayesian odds factors. The second computer algorithm, HADES (Hamiltonian Density Estimation and Sampling), is specifically designed to infer the fully evolved cosmic density field deep into the non-linear regime. In particular, HADES accurately treats the non-linear relationship between the observed galaxy distribution and the underlying continuous density field by correctly accounting for the Poissonian nature of the observables. This allows for very precise recovery of the density field even in sparsely sampled regions. HADES also provides a complete statistical description of the non-linear cosmic density field in the form of a sampled representation of a cosmic density posterior. Beside the possibility of reporting any desired statistical summary of the density field or power-spectrum, such representations of the according posterior distributions also allow for simple non-linear and non-Gaussian error propagation to any quantity finally inferred from the analysis results. The application of HADES to the latest Sloan Digital Sky Survey data denotes the first fully Bayesian non-linear density inference conducted so far. The results obtained from this procedure represent the filamentary structure of our cosmic neighborhood in unprecedented accuracy.