Our page check returns a score out of four, and the decision does not use it

Описание и применение

Проверка страниц в Notifio теперь требует наличие цены и либо адреса, либо контактной информации, чтобы считать страницу объявлением, что устраняет ложные положительные ответы.

Задача

The scraper mistakenly replies to non‑listing pages, causing costly false positives.

Результат

False positives dropped by enforcing a rule that requires a price and either an address or contact method, while still computing a diagnostic score.

Рабочий процесс

• 1. Scraper collects candidate pages via anchor sweep.
• 2. Page content is scanned for price, address, and contact signals using regex patterns.
• 3. A score is computed (price=2, address=1, contact=1).
• 4. The rule "price AND (address OR contact)" determines if the page is a listing.
• 5. If the rule fails, the page is refused with a diagnostic score.
• 6. Refusal messages include the score for later diagnostics.

Доказательства

The expensive failure here is not missing a listing. It is sending a well‑formed enquiry about something that is not a listing.

Источник

A price, plus at least one of an address or a way to contact somebody.

Источник

The function that does it looks like a scoring model and then decides with a rule that ignores the score.

Источник