<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body style="overflow-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;">
<span style="font-size: 14px;">Hi everyone,</span>
<div><span style="font-size: 14px;"><br>
</span></div>
<div><span style="font-size: 14px;">The first lunch talk of this term will be given by Rhys, at
<b>1pm Thursday in C14 (main physics building)</b> - note the room change! Title and abstract are below.</span></div>
<div><span style="font-size: 14px;"><br>
</span></div>
<div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; color: rgb(0, 0, 0);">
<span style="font-size: 14px;">TITLE:<b> Identifying group galaxies merging with massive clusters using machine learning.</b></span></div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; color: rgb(0, 0, 0);">
<b style="font-size: 14px;"><br>
</b></div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; color: rgb(0, 0, 0);">
<span style="font-size: 14px;">ABSTRACT: <i>The environment plays a critical role in shaping galaxy evolution. Galaxy clusters and their surroundings offer diverse conditions that influence galaxies before they reach the dense cluster core, known as “pre-processing”.
However, robustly identifying environmental substructures, particularly galaxy groups in the infall regions, remains a significant challenge. Traditional phase space clustering methods struggle in these transitional zones due to severe projection effects and
finger-of-god (FoG) distortions, precisely where distinguishing between local and global environments is most important. In this talk, I will present a supervised machine-learning framework for identifying group galaxies in and around clusters using projected
positions and radial velocities. Our model trains on mock observations derived from cosmological simulations tailored to match survey conditions. It classifies galaxies into three environmental categories: main cluster, group, and neither (field). The model
achieves an overall accuracy of 75% and a class-weighted precision of 81%. The main cluster class is most successfully recovered with a recall of 84%, followed by the group (77%) and neither (70%) classes. Model performance for group classification is notably
suppressed within 1xR200. However, resampling strategies allow users to tune the model for precision or recall, depending on their scientific goals. The model remains unbiased across various cluster masses and dynamical states. Our method is flexible, observationally
motivated, and well-suited for upcoming spectroscopic surveys. It provides a promising avenue for disentangling environmental influences on galaxy evolution across large-scale structures.</i></span></div>
</div>
<div><font size="2"><br>
</font></div>
<div><span style="font-size: 14px;">Thanks,</span></div>
<div><span style="font-size: 14px;">Joe</span></div>
</body>
</html>