Dr. Wycondia West’s Post

BEH 2024 | The diffusion of statistical metanorms: Paths, processes, and limits 📅April 10th 🕒 2.00 pm 📍San Giobbe - Aula Volpato Embark on a journey into the complex world of statistical metanorms. In our next BEH seminar, we will explore the intricate web of paths through which these norms diffuse. Discover more: https://lnkd.in/dZrUHJxe

Freshly published in Physical Review E, our work on the study of "Speed-accuracy trade-offs in best-of-n collective decision making through heterogeneous mean-field modeling"! https://lnkd.in/ef5EaHNr

Minority opinion alert: I don't think Cohen's kappa is the measure one should use to test agreement of different sources of relevance judgments for retrieval test collections. Test collections contain vast numbers of non-relevant documents, and almost all of those documents are easily determined to be not relevant. Since kappa treats all judgments equally, there is a very strong trend for kappa to increase simply by judging a greater proportion of the document set. This imbalance between the size of the relevant and non-relevant sets is also why IR doesn't use ROC curves to measure retrieval performance: the measure is dominated by the vast number of non-relevant documents which is precisely the part we don't care about. Overlap, focuses on the documents we care about, namely the relevant and nearly relevant. Overlap was defined in the paper by Lesk and Salton as the size of the intersection of the relevant sets divided by the size of the union of the relevant sets. (M.E. Lesk and G. Salton. "Relevance Assessments and Retrieval System Evaluation", _Information Storage and Retrieval_, vol. 4, pp. 343--359, 1969.)   

What sample size do you need to collect to conduct a meaningful analysis? A simple answer might be "more," but collecting data costs both money and time. In this blog post, I describe how to calculate the necessary sample size to achieve statistical significance using the Chernoff Bound.

I am thrilled to announce the publication of my first-author paper. The paper, titled "Utility in Time Description in Priority Best–Worst Discrete Choice Models: An Empirical Evaluation Using Flynn’s Data," has been published in the journal 'Stat.' You can access the paper here: Utility in Time Description in Priority Best–Worst Discrete Choice Models.

I am happy to share with you our latest published article titled "Application of gap metric for model selection in linear parameter varying model-based predictive control". It is accessible via https://lnkd.in/e8ugiyZZ

Great article: frequent mistakes people make in interpreting p-values, confidence intervals and statistical power and how to correctly interpret. Article 👇 https://t.co/uq1sdJVeeb

Retrieval-Augmented-Generation (RAG) is still relevant even with the big LLM's context available now. In terms of efficiency, a big context is expensive and not precise enough to retrieve specific chunks of data. Interesting. https://lnkd.in/eFvPHi5D

"Data is information, and information is power." Radi Hilaneh I didn't know there was a Love Your Data Week, but I'm excited about it. I love data, because without it you are left in the dark and can't make informed decisions. As a researcher I actively collect data, but people passively collect data every day. Everything you see, feel, hear, and taste is data that your brain automatically processes and reacts to. Sometimes these reactions are subconscious, like laughing when you hear a funny joke. Other times these reactions are a conscious effort, like deciding what to eat for dinner. they are all the result of the collection of data. While I can't attend this year's Data-Themed Cards Against Humanity in-person, you better believe this is something I'm going to look into for my next team building activity. #datadrivendecisions #data

Interested in getting CTD best practices for profiling applications? Let our team walk you through how to get quality data from your Sea-Bird Scientific instruments in this Sea-Bird University session, originally filmed in October of 2023. Watch now 👇 https://bit.ly/3MRlr4z