Table 1.

KDD steps and activities involved in this work.

KDD Process
Pre-KDD	Precision psychiatry using ML algorithms’ principal objectives of treatment response analysis, early identification, suicide prevention, real-time monitoring, and subclassified actual mental disorders [33]. In addition, ML models avoid generic diagnoses, providing new classifications of individuals by their features [40].
Selection	The Depresjon and Psykose datasets contain monitor-activity counts of patients with depression and schizophrenia, respectively.
Preprocessing	All patients’ activity count data are concatenated into a single matrix, standardized, transposed, and grouped by hours.
Transformation	After hourly segmentation, data are grouped into subsets following the day stage: morning (06:00–11:59), afternoon (12:00–17:59), evening (18:00–23:59), and night (00:00–05:59).
Data Mining	Classification of depressive, schizophrenic, and control episodes is performed with a random forest classifier.
Interpretation/evaluation	Precision, recall, F1 score, MCC, and accuracy measure every model’s effectiveness to identify healthy, schizophrenic, and depressive episodes concerning the day stage.
Post-KDD	It is not limited to this written report.