Tim Raupach’s Post

I recently posted about my researches on non-linear shrinkage of weighted sample covariance matrices, and here is the second work on this topic. It addresses the problem of population spectrum retrieval and sample spectrum density estimation through a new algorithm, WeSpeR, leveraging recent results in Random Matric Theory. As always, thanks to Gabriel Turinici and Alexandre Miot for their support. The preprint (under review) is available at: https://lnkd.in/eZVRGFrm

Check out the preprint for our new manuscript, led by Dr. Zhe Feng (PNNL), that explores how well characteristics of Mesoscale Convective Systems (MCSs) are simulated by different global storm-resolving models. We ran eight different tracking algorithms to quantify how sensitive our conclusions are to the method of MCS detection and tracking: https://lnkd.in/gfTvammT

Publication alert!! Multiscale and Multidisciplinary Modeling, Experiments and Design Journal (IF - 2.2, Q-3) Title - Assessment of the uniaxial compressive strength of intact rocks: an extended comparison between machine and advanced machine learning models Keywords - Intact rock, Uniaxial compressive strength, Gaussian process regression, Multicollinearity impact https://lnkd.in/dFMZ2cnn

Read about FHWA's efforts in developing Realistic Artificial Data (RAD). Sponsored by FHWA's Exploratory Advance Research (EAR) program, researchers created a synthetic data generation framework on macroscopic and microscopic levels for various roadway facility types and developed a web-based tool. The tool allows users to generate RADs for multiple years at macroscopic segment and microscopic trip levels that could be used to develop crash modification factors or functions and statistical models to determine how the models best represent real-world relationships. https://lnkd.in/eya4WwtM

Publication alert! 💡 I am happy to share that our new paper (joint with Lukas Trottner) on multivariate change estimation for a stochastic heat equation is now available. Our results may, for instance, be relevant in material sciences and image estimation. We studied a stochastic heat equation in arbitrary dimension with discontinuous diffusivity, and we successfully estimated both its jump location and the corresponding distinct diffusivity values. Read more here: https://lnkd.in/d8Xm9r8U

I am happy to share that our research paper titled "Predicting Temporal Clear Water Scour Depth Around Bridge Piers with XGBoost and SVM–PSO Approaches" has been published in the Journal of Hydroinformatics by IWA (SCIE, IF: 2.2)! 🎉 In this study, we explored the critical issue of scouring around bridge piers, a phenomenon that can lead to significant structural failures if not properly understood and managed. By utilizing advanced machine learning techniques like eXtreme Gradient Boosting (XGBoost) and Support Vector Machine (SVM) with Particle Swarm Optimization (PSO), we developed models that accurately predict clear water scour depth under various conditions. I would like to express my gratitude to my co-authors, Prince Gaurav, Lohit Reddy, Dr. Bhabani Shankar Das, and Dr. Balaji Naik, for their collaboration and support throughout this research work. 🔗 #https://lnkd.in/d-8r5Tm5 Let’s continue to push the boundaries of civil engineering and ensure the safety and longevity of our infrastructure! #Research #CivilEngineering #MachineLearning #Hydroinformatics #BridgeSafety #ScourDepth #XGBoost #SVM_PSO #NITPatna

I am proud to share that our latest (and my first!) contribution entitled "A nonparametric penalized likelihood approach to density estimation of space-time point patterns" has been published in "Spatial Statistics". 💡 We propose a novel nonparametric method to estimate the unknown spatio-temporal probability density function associated with point patterns spatially observed on complex domains of various kinds. ✏ We establish some important theoretical properties of the considered estimator and develop a flexible and efficient estimation procedure. 📊 We thoroughly validate the proposed method, by means of several simulation studies and applications to real-world data. Authors: Blerta Begu, Simone Panzeri, Eleonora Arnone, Michelle Carey, Laura M. Sangalli Code available at: https://lnkd.in/eX9KsgqQ Paper available at: https://lnkd.in/eqKQ_mzz #nonparametric #densityestimation #spatiotemporal #pointpatterns

In this research, we proposed moment-based approximations to some stationary characteristics of the process X(t), which describes a semi-Markovian inventory model of type (s,S).

I would like to announce that our pre-print entitled “The role of interface boundary conditions and sampling strategies for Schwarz-based coupling of projection-based reduced order models”, coauthored with Chris Wentland, Francesco Rizzi and Joshua Barnett, is available on ArXiv: https://lnkd.in/gR_Pgyua. The paper demonstrates that for cell-centered finite volume discretizations and non-overlapping domain decomposition, it is possible to obtain a stable and accurate coupled model using Dirichlet-Dirichlet (rather than Robin-Robin or alternating Dirichlet-Neumann) transmission conditions on the Schwarz boundaries. We additionally explore the impact of boundary sampling when utilizing the Schwarz alternating method to couple subdomain-local hyper-reduced PROMs. Please feel free to share with anyone who might be interested.