The pattern people mean by this name is the beginner who has learned enough to form opinions but not enough to know their limits. In trading it has a characteristic timeline: a first profitable month, followed by rapid size increases, followed by an education.
The research itself is narrower and much argued over than the meme suggests, so use the term carefully. The usable core is real, though: self-assessment is worst where knowledge is thinnest, and markets pay out randomly enough in the short run to confirm any belief.
The remedy is a measured sample-size before conclusions, and keeping a written prediction record so your confidence can be scored rather than felt. See confidence-calibration.
Related: overconfidence, optimism-bias, sample-size