2025 Annual Meeting

Every year, the Scientific Programming Committee (SPC) for The Classification Society reviews all of the abstracts submitted for an oral presentation and selects up to 9 contributors to give oral presentations at the Annual Meeting based primarily on the novelty and topic variety of the submissions. This is a competitive process where the success rate depends on the number of submissions received, but in recent history has sat around 50-65%. Unsuccessful submissions are generally encouraged to present their work during the poster session. The SPC decisions are final, and the committee reserves the right to balance the program with respect to additional considerations - for example, the academic level of speakers, variety of topics, etc.

Submission details

For the 2025 annual meeting, the SPC has decided to ask each contributor applying for an oral presentation for more details of their proposed talk. On the abstract submission form, you will find text entry boxes requesting the following details:

1. Short biography (max 500 characters). This information will be used by the chair of the session to introduce the speaker.
2. Title (max 50 characters). This information will be provided in the program.
3. Abstract (max 1200 characters). This information will be provided in the program.
4. Motivation and objectives (max 1000 characters). Describe the motivation(s) for, and objective(s) of, the proposed work. For example, what scientific problem did classification methods help you solve? Or, why did you develop the proposed statistical model? What drawbacks in the current literature are you trying to address? What are the applications of interest that need this new tool? Explain why this project is important.
   This will only be used during committee deliberations and will not appear publicly anywhere.
5. Abbreviated literature review (max 1200 characters). Based on the motivation(s) discussed previously, provide up to 3 full citations and brief context for them with respect to your talk. For example, what are some established approaches to your scientific problem? What are the main competitors to your new method?
   This will only be used during committee deliberations and will not appear publicly anywhere.
6. Key findings and future work (max 1000 characters). Describe the key findings of your work. For example, does the proposed model outperform other state-of-the-art models using simulations or real data sets? Do classification methods help answer important scientific problems in your field? What will future work in the area look like? Here, we are looking for you to explain the most important results from your proposed project and provide forward-looking context.
   This will only be used during committee deliberations and will not appear publicly anywhere.

Important notes:

1. Contributors applying for a poster presentation are only required to fill out the first three text boxes.
2. Boxes 4 - 6 will only appear for those who select 'Oral' under Preferred Presentation Style.

Example

1. Short biography
   Janet Dough is The Classification Society's Professor of Interdisciplinary Data Science. They are the North American Research Chair of Awesomeness.
2. Title
   What were they thinking? New thoughts on variable selection
3. Abstract
   We begin by reviewing several variable selection techniques in the context of clustering and classification and point out several flaws in the existing methodologies. We develop an alternate approach via pure magic and show how pure magic is capable of handling literally any type of data and scientific problem in existence. We compare our perfect results to the existing junk in the literature.
4. Motivation and objectives
   We noticed that the work by Smith (2014) was fundamentally flawed since it avoided any assumptions surrounding pure magic (Rader, 2018). We know that pure magic is capable of handling any problem if you just believe (Subbotsky, 2014). Belief in magic is underused in the unsupervised literature. We wanted to illustrate its utility in feature selection within the classification context.
5. Abbreviated literature review
   Flawed approach that we provide an alternative methodology for:
   Smith, J. (2014). Ultimate Variable Selection. Journal of Selecting Variables, 37(6), 122-123.
   Theoretical basis of our proposal:
   Rader, R. (2018). It's Magic, Pure and Simple. The Exceptional Parent, 48(11), 4-6.
   Rationale for applying the theory within our field:
   Subbotsky, E. (2014). The belief in magic in the age of science. Sage Open, 4(1), 2158244014521433.
6. Key findings and future work
   We found that pure magic was a useful assumption to fix literally every conceivable methodological flaw in the existing literature surrounding feature selection in clustering/classification. We show this on several common data sets (crabs, wine, etc) and extensive simulation studies. We think there is probably a similar utility for pure magic in the context of linear regression, and we will discuss this briefly in the presentation, though we will largely leave it for future consideration.

The Classification Society is pleased to announce that it will continue to support students interested in presenting at the Annual Meeting. At the 2025 annual meeting, every student presentation will be considered for an oral or poster presentation award. These awards come with monetary prizes of at most 500 USD for best oral presentation and at most 250 USD for best poster presentation. Further, any student who gives either an oral or poster presentation at the 2025 annual meeting will also have their registration fee covered by The Classification Society.