# US 2019/0171675 A1 · Advanced Search-Term Disambiguation

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

- Number: US 2019/0171675 A1
- Document ID: US20190171675A1
- Status: Published application
- Kind: A1
- Application number: 16/204,713
- Filed: 2018-11-29
- Published: 2019-06-06
- Earliest claimed: 2017-03-31
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Alexandru Pulbere; Craig A. Statchuk; Changying Sun
- Web page: https://patentgrep.com/patent/US20190171675A1

## CPC

- G06F 16/9538: Presentation of query results
- G06F 16/24578: {using ranking}
- G06F 16/334: Query execution (filtering based on additional data G06F16/335)
- G06F 16/90324: {using system suggestions}
- G06F 16/951: Indexing; Web crawling techniques
- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A mechanism is provided for advanced search-term disambiguation. Responsive to detecting a search term being entered into an autocomplete search field of a search engine, a determination is made of a set of terms from a storage device upon which a search is to be performed. For each term in the set of terms, a determination is made of a semantic distance and independence (Sdi) score to each term candidate in a set of term candidates. The results of the semantic distance and independence (Sdi) scores are ranked and then pruned down to a predetermined number of autocomplete results. A subset of term candidates associated with the predetermined number of autocomplete results from the set of term candidates is then presented to a user as autocomplete suggestions to the search.

## Claims

### Claim 1

A method, in a data processing system, for advanced search-term disambiguation, the method comprising:
responsive to detecting a search term being entered into an autocomplete search field of a search engine, determining a set of terms from a storage device upon which a search is to be performed;
for each term in the set of terms, determining a semantic distance and independence (Sdi) score to each term candidate in a set of term candidates;
ranking results of the semantic distance and independence (Sdi) scores;
pruning the ranked results down to a predetermined number of autocomplete results; and
presenting a subset of term candidates associated with the predetermined number of autocomplete results from the set of term candidates to a user as autocomplete suggestions to the search.

### Claim 2 (depends on 1)

The method of claim 1, wherein the semantic distance and independence (Sdi) score is determined using the following function:
Sdi=(D(t,aj)*O(t,aj))
where D(t,aj) which is a function that measures semantic distance (measure of opposite terms) for given term t and potential term candidate to be offered aj multiplied by O(t,aj) which is a function that measures semantic orthogonality (most independent) for given term t and potential term candidate to be offered aj.

### Claim 3 (depends on 2)

The method of claim 2, wherein the semantic distance is a distance score (i.e. number of hops) from an occurrence of the term to a highest hierarchical parent associated with the term.

### Claim 4 (depends on 2)

The method of claim 2, wherein the semantic orthogonality is an orthogonality score (i.e. a number of independent hops) from each highest hierarchical parent one or more entities identified as having a relationship to the search term.

### Claim 5 (depends on 1)

The method of claim 1, wherein the storage device is an enterprise storage device.

### Claim 6 (depends on 1)

The method of claim 1, wherein the subset of terms candidate are most different terms associated with the search term in the storage device as identified by semantic distance and semantic orthogonality.

### Claim 7 (depends on 1)

The method of claim 1, wherein the predetermined number of autocomplete results are at least one of user identified, automatically determined based on a ratio associated with identified candidate terms, or automatically determined based on a number of candidate terms above an identified threshold.

### Claim at position 8

8-20. (canceled)
