I have a dataset of 300 observations, of which 200 are normal, and the rest have the disease. I have the cognitive assessment scores of these 300 participants, and the assessment is divided into different sections: delusions, depression, anxiety, etc. I'm wondering what technique(s) would be good to look at the variation within the data. And whether a certain section's score would be a good predictor of the participant having the disease. I know of logistic regression, and with that I can construct a classification tree and an ROC curve. Are there any other techniques that I can use here? I'm open to different approaches (techniques for predicting, or just any other ways to look at the data in a meaningful way).
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Based on your description you could use a plethora of methods, any classification methods would do (Random Forest, PLS-DA, LDA, QDA, SVM), as long as your model is validated.
You might also want to look in ANOVA like methods (or rather MANOVA). if the difference between patients and controls is really large exploratory methods like PCa would probably also show you something useful
