SETS: Leveraging Self-Verification and Self-Correction for Improved Test-Time Scaling