BookMatcher

Using TF-IDF (Term Frequency-Inverse Document Frequency) and cosine similarity, I can more accurately find the most matching folder for a URL. Specifically:

Analyze local folders to generate a TF-IDF dictionary

Monitor bookmark save operations, extract keywords from the URL, and find the best matching folders in the TF-IDF dictionary

Display folders ranked by matching score from highest to lowest

This approach has several advantages:

The more pages saved in a folder, the higher the cosine similarity, meaning matches become more accurate

Completely local computation, no need for GPT or other AI tools, with millisecond-level calculation speed