Galaxy NGC 4254 lies 1.5 million light-years from the center of the Virgo cluster. It drags a long stream of cooling hydrogen — a sign that it's plowing headlong into the hot gas. Losing cold fuel quenches the birth of new stars. But the tail alone doesn't tell whether this is a first plunge into the hot medium or a galaxy that has already passed the core and is coming back.

Galaxy clusters are true crucibles. In their centers, gas is swept out and star formation is suppressed. Outskirts still host blue spirals, but among them lurk 'returners'—galaxies that have shot through the core and been flung back by gravity. They look like newcomers, though they have already lost gas and part of their dark halo. Blending these two evolutionary paths muddles the picture: without knowing the orbit, it's unclear where the environment struck — in a group before infall or deep inside the cluster.

To separate newcomers from returners, astrophysicists turned to Machine Learning. They trained a random forest classifier on 324 detailed hydrodynamic simulations of clusters (the Three Hundred project). Each virtual galaxy was described by 14 observable parameters: distance to the center, line-of-sight velocity (from spectroscopy), magnitude gap of brightest members, sky motion direction, and distance to the fourth neighbor. The model doesn't need a full dataset: learning the patterns, it builds a forest of decisions from whatever features the user supplies. The probability threshold can be slid, trading purity for completeness.

Trained on the full feature set, the forest yields an overall accuracy of 84%. Returners are found with about 75% precision and recall, newcomers with 88%. For comparison, traditional methods with hard phase-space boundaries at 61% purity find only 40% of returners. Even with just three basic quantities (distance, line-of-sight velocity, and magnitude gap) the model achieves 74% accuracy. Adding sky velocity direction raises the bar to 81%.

Especially telling is the method's performance on the Virgo cluster. Ten galaxies with stripped hydrogen tails in the outer corona, for which motion directions could be measured, are classified as newcomers. That means the cluster's hot gas strips cold gas almost immediately upon entry, leaving little after a single pass through the crucible.

This tool changes how we study galactic evolution. No longer must we guess what snuffed out star formation on the outskirts — pre-processing in a group or a burn in the core. Clean samples will let us reconstruct the sequence of spirals turning into lenticulars. And so NGC 4254 and its neighbors are not scorched wanderers; they are debutantes, their first date with the crucible just beginning.