phishing_dataset Dataset Overview

Dataset Composition

• Collected 500 phishing sites sourced from PhishTank.
• Collected 500 legitimate sites sourced from Alexa.
• The dataset is split with a 70%/30% train‑test ratio.

Feature Description

URL Features

1. Domain similarity: Similarity between the accessed website's domain and the domain of URLs obtained from Alexa or PhishTank, computed using the Ratcliff‑Obershelp algorithm.
2. URL length: Number of characters in the URL.
3. HTTP protocol type: Standard (0) or secure (1).
4. Number of '.' characters: Count of dot symbols in the URL.
5. Number of '/' characters: Count of slash symbols in the URL.
6. Number of '//' sequences: Count of double‑slash symbols in the URL.
7. Number of '-' characters: Count of hyphen symbols.
8. Number of '_' characters: Count of underscore symbols.
9. Number of '=' characters: Count of equal signs.
10. Number of '(' and ')' characters: Count of parentheses.
11. Number of '{' and '}' characters: Count of curly braces.
12. Number of '[' and ']' characters: Count of square brackets.
13. Number of '<' and '>' characters: Count of angle brackets.
14. Number of '~' characters: Count of tilde symbols.
15. Number of '*' characters: Count of asterisks.
16. Number of '+' characters: Count of plus signs.
17. Presence of '@' symbol: Whether the URL contains '@' (1 = yes, 0 = no).
18. Presence of IP address: Whether the URL contains an IP address (1 = yes, 0 = no).

HTML Features

19. Number of tags: Count of tags used to create hyperlinks or anchor links.
20. Number of tags: Count of tags used for various form elements.
21. Number of tags: Count of tags for clickable buttons.
22. Number of tags: Count of tags for linking external resources such as stylesheets, icons, etc.
23. Number of : Count of <iFrame> tags for embedding external resources such as other HTML documents, video, etc.