Can you explain the TF-IDF and cosine similarity algorithms you implemented in your Python-based Applicant Tracking System and how they contribute to candidate matching?