Presentations

Home Team Programs Communications Products

IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2021

: Ellipse Detection and Localization With Applications to Knots in Sawn Lumber Images, Shenyi Pan, Shuxian Fan, Samuel W.K. Wong, James V. Zidek and Helge Rhodin

Forest Products Stochastic Modeling Group, 2020 Annual Meeting

: Permutation tests under a rotating sampling plan: a problem in long term monitoring program, Jiahua Chen, Yukun Liu and James V. Zidek
Statistical framework for DOL models and degradation: a review and current progress, Samuel W.K. Wong
New approaches for characterizing the strength of lumber based on a spatial distribution of knots, Shuxian Fan, Samuel W.K. Wong and James V. Zidek
Crop and align scanning images, Xin Ding

Joint Statistical Meetings (JSM), 2018

Presentations:

Stopping time detection in functional time series: An application to wood panel glue curing process, Jiguo Cao, Peijun Sang, Kevin Groves & Martin Feng
Assessing Accumulated Damage in Forest Products, Samuel Wong, Chun-Hao Yang & James V. Zidek
Semiparametric monitoring test based on clustered data, Jiahua Chen, Pengfei Li, Yukun Liu & James V. Zidek

Forest Products Stochastic Modeling Group, 2018 Annual Meeting

Log Rotation - Effect of Log Shape, Bruce Lehmann
Bolted Glulam Connection Tests & Effective Stiffness of CLT, Zhiyong Chen
Sparse Functional Partial Least Squares Regression Model, Tianyu Guan, Jiguo Cao, Kevin Groves and Martin Feng
Empirical Likelihood Ratio Test on Quantiles under a Density Ratio Model, Gong (Archer) Zhang and Jiahua Chen
Comparison between Artificial Censored Method and DRM - Empirical Likelihood Method, Xinyao Fan
Enhanced grading systems, Fatema Tuz Jhohura

Joint Statistical Meetings (JSM), 2017

Presentation:

Monitoring Test for One-Sided Multiple Quality Indices, Guangyu Zhu and Jiahua Chen

Posters:

Sequential Graph Matching with Sequential Monte Carlo, Seong-Hwan Jun, Samuel W.K. Wong, James V. Zidek & Alexandre Bouchard-Côté
Sequential Computer Experiments for Failure Probability Estimation, Hao Chen & William J. Welch

First International Young Scholars Forum of Nanjing University of Posts and Telecommunications. 2017

Gaussian Process and Its Application in Engineering, Hao Chen

Abstract

Stresses affect strength properties of a lumber specimen simultaneously. A strong stochastic dependence between these properties would diminish a wooden structure's reliability. Assessing strength relationships is a seemingly insurmountable challenge as every one of these strengths needs a destructive testing to measure. Here is where statistics comes to the rescue. The paper describes a large-scale experiment designed and conducted by Statistics graduate students in a commercial testing lab. It proposes the new theory to infer the stochastic dependence, which applies proof load technique with a penalizing likelihood approach motivated by a Bayesian conjugate prior. The result: asymptotically consistent, more efficient.

International Conference on the Interface between Statistics and Engineering (ICISE), 2014 City University of Hong Kong

Sequential Computer Experiments for Quantile Estimation, Hao Chen & William J. Welch

Statistical Society of Canada (SSC), 2014 Annual Meeting

Presentation:

Evaluating the Agreement Between Two Veneer Thickness Measuring Tools, Guohai Zhou, Brad Wang, Hongbin Zhang & Lang Wu

Posters:

Sequential Experimental Design for Percentile Estimation - A Floor System Example, Hao Chen & William J. Welch

Abstract

Computer models are often used to study physical and engineering processes. In this poster, we first introduce a computer (numerical) model implemented as part of the ongoing work of the Forest Products Stochastic Modelling Group at UBC. It is able to quantify the relationship between Modulus of Elasticity (MOE) and the corresponding deflection under static load in floor systems. We then build a statistical surrogate for the input-output relationship of the numerical model with a modest number of evaluations. We apply sequential methodology to the surrogate for estimating the extreme percentile, say 97.5%, of the deflection distribution. We find that the accuracy of the surrogate is improved with a few more runs chosen by the sequential design.

Quality Over Quantity: Damage Accumulation Models for Lumber Products, Yanling Cai & James V. Zidek

Abstract

Lumber, as a natural building material, affects physical and mechanical properties of the structures built from it. Proof loading is a quality control technique in the lumber industry, which tests if a specimen survives breakage up to a predetermined load level. The survivors are later distributed to the market. However, a large-scale experiment designed and conducted by statistics graduate students in a commercial testing lab indicates that the survivors are damaged. The quality control concerns the severity of the damage accumulated during proof loading. This poster explores potential statistical damage models driven by the data, and compares them to the damage accumulation models serving as a basis of the National Design Specification for Wood Construction by the American Wood Council.