# US 2025/0156467 A1 · Systems, methods, and graphical user interfaces for predicting and analyzing action likelihood

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

- Number: US 2025/0156467 A1
- Document ID: US20250156467A1
- Status: Published application
- Kind: A1
- Application number: 18/812,637
- Filed: 2024-08-22
- Published: 2025-05-15
- Earliest claimed: 2023-11-15
- Assignees: SAS Institute Inc.
- Applicants: SAS Institute Inc.
- Inventors: Teresa S. Jade; Julia Moreno; Ashley Mary Beck
- Web page: https://patentgrep.com/patent/US20250156467A1

## CPC

- G06F 16/35: Clustering; Classification
- G06F 16/34: Browsing; Visualisation therefor (browsing or visualisation for clustering or classification G06F16/358)
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 40/30: Semantic analysis
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A computer-implemented system, computer-implemented method, and computer-program product includes obtaining a text document that includes text describing an action; extracting one or more action tokens from the text document; executing a plurality of linguistic pattern searches that search the text document for one or more likelihood tokens associated with the one or more action tokens; classifying the action to a likelihood category associated with a respective linguistic pattern search of the plurality of linguistic pattern searches that identified the one or more likelihood tokens; classifying the text document to a respective domain; computing a priority value of the action described in the text document based on an input of the likelihood category and the respective domain; and generating a priority summary artifact that visually prioritizes the text document over one or more other text documents when the priority value of the action satisfies a predefined maximum priority threshold value.

## Claims

### Claim 1

A computer-program product comprising a non-transitory machine-readable storage medium storing computer instructions that, when executed by one or more processors, perform operations comprising:
obtaining a text document that includes text describing an action performable in the real world by a person associated with the text;
extracting, via a text processing model, one or more action tokens that represent the action described in the text document;
executing, via the text processing model, a plurality of linguistic pattern searches that search the text document for one or more likelihood tokens associated with the one or more action tokens;
classifying, via the text processing model, the action to a likelihood category associated with a respective linguistic pattern search of the plurality of linguistic pattern searches that identified the one or more likelihood tokens, wherein the likelihood category indicates a likelihood of the action taking place in the real world;
classifying, via a domain classification model, the text document to a respective domain;
computing, via an activity prioritization model, a priority value of the action described in the text document based on an input of the likelihood category and the respective domain; and
generating a priority summary artifact that visually prioritizes the text document over one or more other text documents when the priority value of the action satisfies a pre-defined maximum priority threshold value.

### Claim 2 (depends on 1)

The computer-program product according to claim 1, wherein:
the likelihood of the action represents a probability of the action being performed by the person in the real world, and
the activity prioritization model computes the priority value of the action by:
determining a severity value of the likelihood category,
determining a severity value of the respective domain, and
calculating the priority value of the action using a weighted sum calculation that applies a first weight value to the severity value of the respective domain and a second weight value to the severity value of the likelihood category.

### Claim 3 (depends on 1)

The computer-program product according to claim 1, wherein:
the person associated with the text corresponds to an author of the text document,
the author of the text document describes and performs the action,
the text processing model groups the plurality of linguistic pattern searches into a plurality of likelihood categories, including a low likelihood category, an intermediate likelihood category, and a high likelihood category, and
classifying the action to the likelihood category includes:
determining if the respective linguistic pattern search that identified the one or more likelihood tokens is grouped into the low likelihood category, the intermediate likelihood category, or the high likelihood category,
classifying the action to the low likelihood category when the respective linguistic pattern search is determined to be grouped into the low likelihood category,
classifying the action to the intermediate likelihood category when the respective linguistic pattern search is determined to be grouped into the intermediate likelihood category, and
classifying the action to the high likelihood category if the respective linguistic pattern search is determined to be grouped into the high likelihood category.

### Claim 4 (depends on 1)

The computer-program product according to claim 1, wherein:
the plurality of linguistic pattern searches search for linguistic cues that indicate an intent of the person across different timeframes,
searching the text document for the one or more likelihood tokens includes searching the text document for one or more low likelihood tokens, and
searching the text document for the one or more low likelihood tokens at least includes:
searching the text document for one or more tokens associated with a counterfactual conditional;
searching the text document for one or more tokens associated with a subjunctive expression; or
searching the text document for one or more tokens associated with a present-time hypothetical condition.

### Claim 5 (depends on 1)

The computer-program product according to claim 1, wherein:
the action has been planned by the person,
searching the text document for the one or more likelihood tokens includes searching the text document for one or more intermediate likelihood tokens, and
searching the text document for the one or more intermediate likelihood tokens at least includes:
searching the text document for one or more tokens associated with a future-time hypothetical condition;
searching the text document for one or more tokens associated with a future-time predictive conditional;
searching the text document for one or more tokens associated with an anticipatory expression; or
searching the text document for one or more tokens associated with a modal expression.

### Claim 6 (depends on 1)

The computer-program product according to claim 1, wherein:
the likelihood category indicates the likelihood of the action taking place in the real world from a perspective of an author of the text document,
searching the text document for the one or more likelihood tokens includes searching the text document for one or more high likelihood tokens, and
searching the text document for the one or more high likelihood tokens at least includes:
searching the text document for one or more tokens associated with an expression of intent;
searching the text document for one or more tokens associated with a past-time expression;
searching the text document for one or more tokens associated with a present-time expression; or
searching the text document for one or more tokens associated with a future-time expression.

### Claim 7 (depends on 1)

The computer-program product according to claim 1, wherein:
the action is a physical action, and
the respective linguistic pattern search identifies the one or more likelihood tokens by:
searching a sentence of the text document for one or more tokens that satisfy pre-defined token likelihood search criteria.

### Claim 8 (depends on 1)

The computer-program product according to claim 1, wherein:
the text processing model groups the plurality of linguistic pattern searches into a plurality of likelihood categories,
computing the priority value of the action at least includes converting the likelihood category to a severity value, and
the activity prioritization model converts the likelihood category to the severity value by:
searching for a record corresponding to the likelihood category in a lookup data structure that associates the plurality of likelihood categories with a respective severity value, and
converting the likelihood category to the respective severity value defined in the record that corresponds to the likelihood category.

### Claim 9 (depends on 1)

The computer-program product according to claim 1, wherein:
the domain classification model is configured to classify the text document into a plurality of domains,
computing the priority value of the action at least includes converting the respective domain to a severity value, and
the activity prioritization model converts the respective domain to the severity value by:
searching for a record corresponding to the respective domain in a lookup data structure that associates the plurality of domains with a respective severity value, and
converting the respective domain to the respective severity value defined in the record that corresponds to the respective domain.

### Claim 10 (depends on 1)

The computer-program product according to claim 1, wherein:
extracting the one or more action tokens from the text document includes identifying a modal verb from a respective sentence within the text document,
the plurality of linguistic pattern searches includes a temporal-based hypothetical conditional pattern search, and
searching the text document for the one or more likelihood tokens includes searching for one or more tokens within the respective sentence of the text document that satisfies token likelihood search criteria of the temporal-based hypothetical conditional pattern search, wherein the token likelihood search criteria is satisfied when:
a token within the respective sentence of the text document matches to a pre-defined set of conditional words, and
the token that matched to the pre-defined set of conditional words is detected in a pre-defined position relative to the modal verb.

### Claim 11 (depends on 1)

The computer-program product according to claim 1, wherein the priority summary artifact forgoes visually prioritizing the text document over the one or more other text documents when the priority value of the action does not satisfy the pre-defined maximum priority threshold value.

### Claim 12 (depends on 1)

The computer-program product according to claim 1, wherein:
the text document is a chat conversation,
the action described in the text document relates to a discussion of a physical meeting,
the likelihood category associated with the action is one of: a low likelihood category, an intermediate likelihood category, or a high likelihood category,
the respective domain of the text document corresponds to a child safety domain, and
the priority value of the action represents a severity of the physical meeting.

### Claim 13 (depends on 1)

The computer-program product according to claim 1, wherein:
detecting the one or more action tokens includes detecting one or more verbs in the text document, and
the text processing model uses part-of-speech tagging to detect the one or more verbs in the text document.

### Claim 14 (depends on 1)

The computer-program product according to claim 1, wherein:
the domain classification model is a child safety domain classification model, and
classifying the text document to the respective domain includes classifying the text document to one of a plurality of child safety categories associated with the child safety domain classification model.

### Claim 15 (depends on 1)

The computer-program product according to claim 1, wherein:
the priority value of the action described in the text document satisfies the pre-defined maximum priority threshold value, and
visually prioritizing the text document over the one or more other text documents includes:
displaying an entry corresponding to the text document at a higher prominence than a text document that does not satisfy the pre-defined maximum priority threshold value.

### Claim 16 (depends on 1)

The computer-program product according to claim 1, wherein:
the one or more likelihood tokens includes a first set of likelihood tokens that were identified by the respective linguistic pattern search and a second set of likelihood tokens that were identified by a second respective linguistic pattern search of the plurality of linguistic pattern searches,
the likelihood category of the respective linguistic pattern search is different from a likelihood category associated with the second respective linguistic pattern search, and
classifying the action includes:
receiving disambiguation user input that indicates whether the action described in the text document corresponds to the likelihood category associated with the respective linguistic pattern search or the likelihood category associated with the second respective linguistic pattern search, and
in response to detecting the disambiguation user input:
in accordance with a determination that the disambiguation user input included a selection of the likelihood category associated with the respective linguistic pattern search, classifying the action to the likelihood category associated with the respective linguistic pattern search, and
in accordance with a determination that the disambiguation user input included a selection of the likelihood category associated with the second respective linguistic pattern search, classifying the action to the likelihood category associated with the second respective linguistic pattern search.

### Claim 17 (depends on 1)

The computer-program product according to claim 1, wherein:
the one or more likelihood tokens includes a first set of likelihood tokens that were identified by the respective linguistic pattern search and a second set of likelihood tokens that were identified by a second respective linguistic pattern search of the plurality of linguistic pattern searches,
the likelihood category of the respective linguistic pattern search is different from a likelihood category associated with the second respective linguistic pattern search, and
classifying the action includes:
determining if the likelihood category of the respective linguistic pattern search indicates a higher likelihood than the likelihood category associated with the second respective linguistic pattern search,
classifying the action to the likelihood category associated with the respective linguistic pattern search if the likelihood category of the respective linguistic pattern search is determined to have the higher likelihood, and
classifying the action to the likelihood category associated with the second respective linguistic pattern search if the likelihood category of the respective linguistic pattern search is determined to not have the higher likelihood.

### Claim 18

A computer-implemented method comprising:
obtaining a text document that includes text describing an action performable in the real world by a person associated with the text;
extracting, via a text processing model, one or more action tokens that represent the action described in the text document;
executing, via the text processing model, a plurality of linguistic pattern searches that search the text document for one or more likelihood tokens associated with the one or more action tokens;
classifying, via the text processing model, the action to a likelihood category associated with a respective linguistic pattern search of the plurality of linguistic pattern searches that identified the one or more likelihood tokens, wherein the likelihood category indicates a likelihood of the action taking place in the real world;
classifying, via a domain classification model, the text document to a respective domain;
computing, via an activity prioritization model, a priority value of the action described in the text document based on an input of the likelihood category and the respective domain; and
generating a priority summary artifact that visually prioritizes the text document over one or more other text documents when the priority value of the action satisfies a pre-defined maximum priority threshold value.

### Claim 19 (depends on 18)

The computer-implemented method according to claim 18, wherein:
the likelihood of the action represents a probability of the action being performed by the person in the real world, and
the activity prioritization model computes the priority value of the action by:
determining a severity value of the likelihood category,
determining a severity value of the respective domain, and
calculating the priority value of the action using a weighted sum calculation that applies a first weight value to the severity value of the respective domain and a second weight value to the severity value of the likelihood category.

### Claim 20 (depends on 18)

The computer-implemented method according to claim 18, wherein:
the person associated with the text corresponds to an author of the text document,
the author of the text document describes and performs the action,
the text processing model groups the plurality of linguistic pattern searches into a plurality of likelihood categories, including a low likelihood category, an intermediate likelihood category, and a high likelihood category, and
classifying the action to the likelihood category includes:
determining if the respective linguistic pattern search that identified the one or more likelihood tokens is grouped into the low likelihood category, the intermediate likelihood category, or the high likelihood category,
classifying the action to the low likelihood category when the respective linguistic pattern search is determined to be grouped into the low likelihood category,
classifying the action to the intermediate likelihood category when the respective linguistic pattern search is determined to be grouped into the intermediate likelihood category, and
classifying the action to the high likelihood category if the respective linguistic pattern search is determined to be grouped into the high likelihood category.

### Claim 21 (depends on 18)

The computer-implemented method according to claim 18, wherein:
the plurality of linguistic pattern searches search for linguistic cues that indicate an intent of the person across different timeframes,
searching the text document for the one or more likelihood tokens includes searching the text document for one or more low likelihood tokens, and
searching the text document for the one or more low likelihood tokens at least includes:
searching the text document for one or more tokens associated with a counterfactual conditional;
searching the text document for one or more tokens associated with a subjunctive expression; or
searching the text document for one or more tokens associated with a present-time hypothetical condition.

### Claim 22 (depends on 18)

The computer-implemented method according to claim 18, wherein:
the action has been planned by the person,
searching the text document for the one or more likelihood tokens includes searching the text document for one or more intermediate likelihood tokens, and
searching the text document for the one or more intermediate likelihood tokens at least includes:
searching the text document for one or more tokens associated with a future-time hypothetical condition;
searching the text document for one or more tokens associated with a future-time predictive conditional;
searching the text document for one or more tokens associated with an anticipatory expression; or
searching the text document for one or more tokens associated with a modal expression.

### Claim 23 (depends on 18)

The computer-implemented method according to claim 18, wherein:
the likelihood category indicates the likelihood of the action taking place in the real world from a perspective of an author of the text document,
searching the text document for the one or more likelihood tokens includes searching the text document for one or more high likelihood tokens, and
searching the text document for the one or more high likelihood tokens at least includes:
searching the text document for one or more tokens associated with an expression of intent;
searching the text document for one or more tokens associated with a past-time expression;
searching the text document for one or more tokens associated with a present-time expression; or
searching the text document for one or more tokens associated with a future-time expression.

### Claim 24 (depends on 18)

The computer-implemented method according to claim 18, wherein:
the action is a physical action, and
the respective linguistic pattern search identifies the one or more likelihood tokens by:
searching a sentence of the text document for one or more tokens that satisfy pre-defined token likelihood search criteria.

### Claim 25

A computer-implemented system comprising:
one or more processors;
a memory; and
a computer-readable medium operably coupled to the one or more processors, the computer-readable medium having computer-readable instructions stored thereon that, when executed by the one or more processors, cause a computing device to perform operations comprising:
obtaining a text document that includes text describing an action performable in the real world by a person associated with the text;
extracting, via a text processing model, one or more action tokens that represent the action described in the text document;
executing, via the text processing model, a plurality of linguistic pattern searches that search the text document for one or more likelihood tokens associated with the one or more action tokens;
classifying, via the text processing model, the action to a likelihood category associated with a respective linguistic pattern search of the plurality of linguistic pattern searches that identified the one or more likelihood tokens, wherein the likelihood category indicates a likelihood of the action taking place in the real world;
classifying, via a domain classification model, the text document to a respective domain;
computing, via an activity prioritization model, a priority value of the action described in the text document based on an input of the likelihood category and the respective domain; and
generating a priority summary artifact that visually prioritizes the text document over one or more other text documents when the priority value of the action satisfies a pre-defined maximum priority threshold value.

### Claim 26 (depends on 25)

The computer-implemented system according to claim 25, wherein:
the text processing model groups the plurality of linguistic pattern searches into a plurality of likelihood categories,
computing the priority value of the action at least includes converting the likelihood category to a severity value, and
the activity prioritization model converts the likelihood category to the severity value by:
searching for a record corresponding to the likelihood category in a lookup data structure that associates the plurality of likelihood categories with a respective severity value, and
converting the likelihood category to the respective severity value defined in the record that corresponds to the likelihood category.

### Claim 27 (depends on 25)

The computer-implemented system according to claim 25, wherein:
the domain classification model is configured to classify the text document into a plurality of domains,
computing the priority value of the action at least includes converting the respective domain to a severity value, and
the activity prioritization model converts the respective domain to the severity value by:
searching for a record corresponding to the respective domain in a lookup data structure that associates the plurality of domains with a respective severity value, and
converting the respective domain to the respective severity value defined in the record that corresponds to the respective domain.

### Claim 28 (depends on 25)

The computer-implemented system according to claim 25, wherein:
extracting the one or more action tokens from the text document includes identifying a modal verb from a respective sentence within the text document,
the plurality of linguistic pattern searches includes a temporal-based hypothetical conditional pattern search, and
searching the text document for the one or more likelihood tokens includes searching for one or more tokens within the respective sentence of the text document that satisfies token likelihood search criteria of the temporal-based hypothetical conditional pattern search, wherein the token likelihood search criteria is satisfied when:
a token within the respective sentence of the text document matches to a pre-defined set of conditional words, and
the token that matched to the pre-defined set of conditional words is detected in a pre-defined position relative to the modal verb.

### Claim 29 (depends on 25)

The computer-implemented system according to claim 25, wherein the priority summary artifact forgoes visually prioritizing the text document over the one or more other text documents when the priority value of the action does not satisfy the pre-defined maximum priority threshold value.

### Claim 30 (depends on 25)

The computer-implemented system according to claim 25, wherein:
the text document is a chat conversation,
the action described in the text document relates to a discussion of a physical meeting,
the likelihood category associated with the action is one of: a low likelihood category, an intermediate likelihood category, or a high likelihood category,
the respective domain of the text document corresponds to a child safety domain, and
the priority value of the action represents a severity of the physical meeting.
