Integrative diagnosis of psychiatric conditions using ChatGPT and fMRI data

BMC Psychiatry

Table 3 Integrated experiment results (Part 1)

Experiment	Accuracy	Precision	Recall
Baseline Model (without NLP)	0.79	0.76	0.78
Full Model (with NLP)	0.87	0.84	0.85
Ablation Study 1 (Removing NLP Features)	0.81	0.78	0.80
Ablation Study 2 (NLP Features Only)	0.83	0.80	0.82

Accuracy, precision, and recall are compared across different configurations, including removal of NLP features and use of NLP features alone

ISSN: 1471-244X