PSYC 212 Lecture Notes - Lecture 4: Detection Theory, Ovarian Cancer, Biomarker

• School: McGill University
• Department: Psychology
• Course: PSYC 212
• Professor: Mathieu Roy

Document Summary

Also sometimes just by change there is a possibility of imaging that a signal exists. When i have this level of appearance of the square, im gonna say yes. The reason why we need sdt, is to allow us to measure the sensitivity regardless of the criterion set by the person. We try to have a more robust and generalizable method of sensitivity. They have to take a decision whether the person has cancer or not. We are looking to have as much hits as possible. Setting the biomarker at 10, over 10 there is cancer and below no. Try out randomly different criterion and thus we get a curve the best criterion and a curve that resemble the ideal curve. Allows to generalize and get a sensitivity index which will permit to get more hits and less misses. Sensitivity is approximately equal to the difference between hit and false alarms.