The WA Branch of the Statistical Society of Australia are delighted to host an evening dedicated to Early Career and Student Statisticians. All visitors are welcome to attend this event, especially anyone studying or interested in pursuing a career in statistics.

Date: Tuesday, 9 May 2023
Time: 5:30PM - 7:00PM (Livestream starts at 6:00PM)
Location: Cheryl Praeger Lecture Room, The University of Western Australia
Cost: Free. Drinks and pizza sponsored by the Branch.

Two speakers have been invited to speak at this event: Sophie Giraudo (recipient of the 2022 Honours Scholarship) and Tim Pesch (PhD student at UWA).

Presentations

Detection of Lymph Node Metastasis in Endometrial Cancer Patients

Sophie Giraudo

Endometrial cancer (EC) is the most common gynaecological malignancy in Australia, and lymph node metastasis (LNM) is an important prognostic factor. Current staging treatment involves the removal of pelvic lymph nodes, which results in significant complications in 38% of patients. Further, this treatment is unnecessary for 80% of patients. Several statistical and machine learning analyses have been carried out to detect LNM pre-surgery, however, none have been widely implemented in clinical settings. In this analysis, we work with mass spectrometry imaging data of tumour tissue samples from 173 patients with EC and use linear discriminant analysis to classify patients’ LNM status. The analysis involves the reduction of the original data of 170,000 masses to 14,500 bins, and a variable selection process based on canonical correlation analysis to rank the discriminating masses. At a patient level, our analysis results in an accuracy of 100%, exceeding previous analyses. These results have the potential for being applicable in decision making processes of individualised patient treatment.

About the Presenter

Sophie Giraudo recently graduated from the University of Western Australia majoring in Data Science and Mathematics and Statistics, with Honours in Mathematics and Statistics. She was the recipient of the 2023 Honours Scholarship from the WA Branch.

Sophie currently works as a Data Impact Officer for the Office of Homelessness at the Department of Communities, and aspires for a career in academia where she can inspire the next generation of female statisticians and can work in biostatistics and applications of statistics to medical research.

Estimation with Extended Sequential Order Statistics

Tim Pesch

Historically the lifetimes of individual units of multi-component systems are modelled as independent and identically distributed. The introduction of Sequential Order Statistics by Kamps (1995) relaxes this assumption by allowing component lifetime distributions to change upon failure of another unit. This assumption is justified when the stress on the surviving units increases, and their respective lifetime expectation consequently reduces as components fail successively. Most load-sharing systems will exhibit such a relationship between components. Additionally, the assumption of heterogeneous components is reasonable for many technical systems since components are often either of different type or vary in their functions within the system. Take a series of wash tanks that filter an inflow of contaminated product for example. Each tank experiences a different workload based on its position in the series. Consequently, the tanks have different lifetime expectations while possibly being of the same type. Adopting the assumption of heterogeneous components into the model of Sequential Order Statistics results in the Extended Sequential Order Statistics model, first introduced by Baratnia and Doostparast (2017).

In my talk, I will present a series of exciting inferential results for Extended Sequential Order Statistics. Those include different types of maximum likelihood estimators of the underpinning distribution parameters (and model parameters under the conditional proportional hazard rate approach). Further, I will present a likelihood ratio test to help determine which model is better suited, the Sequential Order Statistics or the Extended Sequential Order Statistics model. These estimation and testing techniques can be used to achieve meaningful results in real life applications. They empower system operators to better predict failure times and make educated decisions towards their maintenance schedule and stock piling strategies.

About the Presenter

Tim has been a PhD student at UWA for the past 2.5 years. During his Masters degree at RWTH Aachen University he worked on Progressive Censoring, competing risks and the stress-strength model. Currently he studies a specific type of order statistics called Extended Sequential Order Statistics which fit into the greater scope of reliability analysis. He is supervised by Adriano Polpo, Edward Cripps (both from UWA) and Erhard Cramer (from RWTH Aachen University, Germany).

Refreshments and Dinner

Members, visitors, and guests are invited to mingle over wine, beer, cider, soft drinks and pizza from 5:30PM. Following the meeting you are invited to dine with fellow attendees at a nearby restaurant. Early career and students will receive $20 off their meal.

Meeting directions

The Cheryl Praeger Lecture Room is located on the ground floor of the Mathematics building at The University of Western Australia. Its entrance is on the northern side of the building. See: UWA Maps, Google Maps.

Parking is free on the UWA Crawley campus after 5:00PM. A convenient place to park is Car Park 18 accessible from Fairway Entry 1.

Remote Viewing Option

For those unable to attend in-person, the presentation will be streamed live over Zoom. Please register on this page to get the connection details.

For further information please contact the WA Branch Secretary (ssa.wa.secretary@gmail.com).