Classification of top-level domain (TLD) websites based on a known website classification

1. A method comprising:

receiving, by a computer system, a website Uniform Resource Locator (URL) that is known to be associated with a particular content classification;

generating, by the computer system, a list of candidate domain names, each including a host name of the website URL and a different top-level domain (TLD) extension based on a predefined TLD list;

determining, by the computer system, for each of the candidate domain names in the list of candidate domain names, whether an Internet Protocol (IP) address of the candidate domain name is equal to an IP address of the website URL;

when a result of said determining is affirmative, then associating the particular content classification with the candidate domain name; and

when the result is negative, then performing a cosine similarity measurement process between information associated with the candidate domain name and information associated with the website URL to determine whether to associate the particular content classification with the candidate domain name.

2. The method of claim 1 , wherein the cosine similarity measurement process further comprises:

retrieving anchor information pertaining to the candidate domain name;

generating a first vector for the candidate domain name based on the retrieved anchor information;

computing a cosine similarity metric between the first vector and a corresponding vector of the web site URL;

determining whether the computed cosine similarity metric is greater than a defined threshold; and

when the computed cosine similarity metric is determined to be greater than the defined threshold, then associating the particular content classification with the candidate domain name.

3. The method of claim 2 , wherein the anchor information comprises one or a combination of:

tags of a web page to which the candidate domain name pertains;

a title of the web page;

metadata of the web page;

hyperlinks of the web page;

content of the web page;

display attributes of the web page;

a structure of the web page; and

a layout of the web page.

4. The method of claim 2 , further comprising:

hashing the first vector to generate a first hash; and

wherein said computing a cosine similarity metric involves comparing the first hash with a hash of the vector of the website URL.

5. The method of claim 1 , wherein each candidate domain name in the list of candidate domain names comprises a fully qualified domain name (FQDN).

6. The method of claim 1 , wherein the particular content classification is indicative of pornographic content.

7. A non-transitory computer-readable storage medium embodying a set of instructions, which when executed by one or more processors of one or more computer systems, causes the one or more processors to perform a method comprising:

receiving a website Uniform Resource Locator (URL) that is known to be associated with a particular content classification;

generating a list of candidate domain names, each including a host name of the website URL and a different top-level domain (TLD) extension based on a predefined TLD list;

determining for each of the candidate domain names in the list of candidate domain names whether an IP address of the candidate domain name is equal to an Internet Protocol (IP) address of the website URL;

when a result of said determining is affirmative, then associating the particular content classification with the candidate domain name; and

when the result is negative, then performing a cosine similarity measurement process between information associated with the candidate domain name and information associate with the website URL to determine whether to associate the particular content classification with the candidate domain name.

8. The non-transitory computer-readable storage medium of claim 7 , wherein the cosine similarity measurement process further comprises:

retrieving anchor information pertaining to the candidate domain name;

generating a first vector for the candidate domain name based on the retrieved anchor information;

computing a cosine similarity metric between the first vector and a corresponding vector of the web site URL;

determining whether the computed cosine similarity metric is greater than a defined threshold; and

when the computed cosine similarity metric is determined to be greater than the defined threshold, then associating the particular content classification with the candidate domain name.

9. The non-transitory computer-readable storage medium of claim 8 , wherein the anchor information comprises one or a combination of:

tags of a web page to which the candidate domain name pertains;

a title of the web page;

metadata of the web page;

hyperlinks of the web page;

content of the web page;

display attributes of the web page;

a structure of the web page; and

a layout of the web page.

10. The non-transitory computer-readable storage medium of claim 8 , wherein the method further comprises:

hashing the first vector to generate a first hash; and

wherein said computing a cosine similarity metric involves comparing the first hash with a hash of the vector of the website URL.

11. The non-transitory computer-readable storage medium of claim 7 , wherein each candidate domain name in the list of candidate domain names comprises a fully qualified domain name (FQDN).

12. The non-transitory computer-readable storage medium of claim 7 , wherein the particular content classification is indicative of pornographic content.