# US 2020/0396204 A1 · Guided word association based domain name detection

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0396204 A1
- Document ID: US20200396204A1
- Status: Published application
- Kind: A1
- Application number: 16/440,150
- Filed: 2019-06-13
- Published: 2020-12-17
- Earliest claimed: 2019-06-13
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Pablo Loyola; Kugamoorthy Gajananan; Yuji Watanabe; Fumiko Akiyama
- Web page: https://patentgrep.com/patent/US20200396204A1

## CPC

- G06F 16/90344: {by using string matching techniques}
- H04L 61/302: {Administrative registration, e.g. for domain names at internet corporation for assigned names and numbers [ICANN]}
- G06F 16/951: Indexing; Web crawling techniques
- G06F 16/9566: {URL specific, e.g. using aliases, detecting broken or misspelled links}
- H04L 61/4511: using domain name system [DNS]
- H04L 2101/30: Types of network names
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Guided word association based domain name detection may be performed by obtaining an original domain name, constructing a feature space from a corpus of text, wherein each word appearing in the corpus is represented as a vector in the feature space, detecting whether a domain name registration exists for each combination of the original domain name and each of a plurality of seed words from the feature space, determining, for each seed word included in an existing domain name registration, a plurality of nearest neighbor candidate words, based on vector distance in the feature space, and repeating, for one or more repetitions, the detecting and the determining, wherein the plurality of nearest neighbor candidate words are utilized as the plurality of seed words.

## Claims

### Claim 1

A computer program product including one or more computer readable storage mediums collectively storing program instructions that are executable by a computer to cause the computer to perform operations comprising:
obtaining an original domain name;
constructing a feature space from a corpus of text, wherein each word appearing in the corpus is represented as a vector in the feature space;
detecting whether a domain name registration exists for each combination of the original domain name and each of a plurality of seed words from the feature space;
determining, for each seed word included in an existing domain name registration, a plurality of nearest neighbor candidate words, based on vector distance in the feature space; and
repeating, for at least one repetition, the detecting and the determining, wherein the plurality of nearest neighbor candidate words are utilized as the plurality of seed words.

### Claim 2 (depends on 1)

The computer program product of claim 1, further comprising disqualifying, from the plurality of nearest neighbor candidate words of each seed word included in an existing domain name registration, each candidate word having a vector distance within a threshold distance of any seed word included in a non-existent domain name registration;
wherein the repeating further includes the disqualifying.

### Claim 3 (depends on 2)

The computer program product of claim 2, wherein the disqualifying includes maintaining, as a candidate word, every nth candidate word having a vector distance within a threshold distance of any seed word included in a non-existent domain name registration.

### Claim 4 (depends on 1)

The computer program product of claim 1, wherein prior to the repeating, the plurality of seed words include words from the corpus that appear most frequently.

### Claim 5 (depends on 1)

The computer program product of claim 1, wherein each combination of the original domain name and each of the plurality of seed words is one of a combination in which the seed word follows the original domain name and a combination in which the seed word precedes the original domain name.

### Claim 6 (depends on 1)

The computer program product of claim 1, wherein the constructing includes
cataloging the unique words in the corpus, and
generating a vector for each unique word in the feature space.

### Claim 7 (depends on 1)

The computer program product of claim 1, further comprising selecting a corpus of text according to one of location and time.

### Claim 8

A computer-implemented method comprising:
obtaining an original domain name;
constructing a feature space from a corpus of text, wherein each word appearing in the corpus is represented as a vector in the feature space;
detecting whether a domain name registration exists for each combination of the original domain name and each of a plurality of seed words from the feature space;
determining, for each seed word included in an existing domain name registration, a plurality of nearest neighbor candidate words, based on vector distance in the feature space; and
repeating, for at least one repetition, the detecting and the determining, wherein the plurality of nearest neighbor candidate words are utilized as the plurality of seed words.

### Claim 9 (depends on 8)

The computer-implemented method of claim 8, further comprising disqualifying, from the plurality of nearest neighbor candidate words of each seed word included in an existing domain name registration, each candidate word having a vector distance within a threshold distance of any seed word included in a non-existent domain name registration;
wherein the repeating further includes the disqualifying.

### Claim 10 (depends on 9)

The computer-implemented method of claim 9, wherein the disqualifying includes maintaining, as a candidate word, every nth candidate word having a vector distance within a threshold distance of any seed word included in a non-existent domain name registration.

### Claim 11 (depends on 8)

The computer-implemented method of claim 8, wherein prior to the repeating, the plurality of seed words include words from the corpus that appear most frequently.

### Claim 12 (depends on 8)

The computer-implemented method of claim 8, wherein each combination of the original domain name and each of the plurality of seed words is one of a combination in which the seed word follows the original domain name and a combination in which the seed word precedes the original domain name.

### Claim 13 (depends on 8)

The computer-implemented method of claim 8, wherein the constructing includes
cataloging the unique words in the corpus, and
generating a vector for each unique word in the feature space.

### Claim 14 (depends on 8)

The computer-implemented method of claim 8, further comprising selecting a corpus of text according to one of location and time.

### Claim 15

An apparatus comprising:
an obtaining section configured to obtain an original domain name;
a constructing section configured to construct a feature space from a corpus of text, wherein each word appearing in the corpus is represented as a vector in the feature space;
a detecting section configured to detect whether a domain name registration exists for each combination of the original domain name and each of a plurality of seed words from the feature space;
a determining section configured to determine, for each seed word included in an existing domain name registration, a plurality of nearest neighbor candidate words, based on vector distance in the feature space; and
a repeating section configured to cause the detecting section and the determining section to repeat, for at least one repetition, their respective functions utilizing the plurality of nearest neighbor candidate words as the plurality of seed words
wherein the obtaining, constructing, detecting, determining, and repeating sections are implemented by a memory device for storing program code and a hardware processor for executing the program code.

### Claim 16 (depends on 15)

The apparatus of claim 15, further comprising a disqualifying section configured to disqualify, from the plurality of nearest neighbor candidate words of each seed word included in an existing domain name registration, each candidate word having a vector distance within a threshold distance of any seed word included in a non-existent domain name registration;
wherein the repeating section further causes the disqualifying section to repeat its function.

### Claim 17 (depends on 16)

The apparatus of claim 16, wherein the disqualifying section includes a maintaining section configured to maintain, as a candidate word, every nth candidate word having a vector distance within a threshold distance of any seed word included in a non-existent domain name registration.

### Claim 18 (depends on 15)

The apparatus of claim 15, wherein prior to the repeating section causing the detecting section and the determining section to repeat, the plurality of seed words include words from the corpus that appear most frequently.

### Claim 19 (depends on 15)

The apparatus of claim 15, wherein each combination of the original domain name and each of the plurality of seed words is one of a combination in which the seed word follows the original domain name and a combination in which the seed word precedes the original domain name.

### Claim 20 (depends on 15)

The apparatus of claim 15, wherein the constructing includes a cataloging section configured to catalog the unique words in the corpus, and generating section configure to generate a vector for each unique word in the feature space.
