# US 2026/0289269 A1 · System

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

- Number: US 2026/0289269 A1
- Document ID: US20260289269A1
- Status: Published application
- Kind: A1
- Application number: 19/566,340
- Filed: 2026-03-13
- Published: 2026-09-24
- Earliest claimed: 2025-03-19
- Assignees: SoftBank Group Corp.
- Applicants: SoftBank Group Corp.
- Inventors: Tomoya GO
- Web page: https://patentgrep.com/patent/US20260289269A1

## CPC

- G06N 3/0475: Generative networks
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06Q 30/02024: {based on temporal or time series analysis, e.g. seasonality or cyclical}
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A system acquires user identification data from a terminal device and associates the user identification data with account identification data of one or more external service platforms. Target item registration data including attribute data and periodicity data for monitored items is stored. Historical transaction records are collected from the external service platforms, preprocessed, and used as learning data to train a time-series machine learning model that outputs prediction data indicating a predicted time interval until a next occurrence for each monitored item. Current condition data from external information sources is analyzed to determine recommendation data including recommended timing and condition parameters. A prompt data structure including the prediction data and recommendation data is provided to a generative neural network model to generate notification text transmitted to the terminal device.; and transmit the sentence to the user through the communication application.

## Claims

### Claim 1

A system comprising:
circuitry configured to:
acquire, via a communication interface coupled to a packet-switched network, user identification data from a terminal device, associate the user identification data with account identification data of one or more external service platforms, and store the associated identification data in an information storage device;
acquire target item registration data from the terminal device, the target item registration data comprising attribute data and periodicity data for one or more monitored items, and store the target item registration data in the information storage device;
collect historical transaction records from the one or more external service platforms via the communication interface, execute preprocessing on the historical transaction records comprising missing-value completion and timestamp normalization to a unified time axis, and store the preprocessed records as learning data in the information storage device;
train a time-series machine learning model using the learning data as input, and generate prediction data by causing the time-series machine learning model to output, for each monitored item, a predicted time interval until a next occurrence;
acquire current condition data from a plurality of external information providing apparatuses via the communication interface, compute an effective value index for each candidate by applying a comparison algorithm to the current condition data, and determine recommendation data comprising a recommended timing and recommended condition parameters based on the prediction data and stored preference parameters;
construct a prompt data structure by embedding the prediction data and the recommendation data into an instruction sequence specifying notification target parameters, detail level, and length constraints, and transmit the prompt data structure to a generative neural network model comprising a transformer-based architecture to acquire generated notification text data; and
transmit the generated notification text data to the terminal device via the communication interface.

### Claim 2 (depends on 1)

The system according to claim 1, wherein the time-series machine learning model comprises a recurrent neural network configured to process sequential input vectors derived from the learning data, and the prediction data comprises a probability distribution over future time intervals for each monitored item.

### Claim 3 (depends on 2)

The system according to claim 2, wherein the circuitry is further configured to retrain the time-series machine learning model periodically by incorporating newly collected historical transaction records into the learning data and executing an incremental training operation.

### Claim 4 (depends on 1)

The system according to claim 1, wherein the preprocessing further comprises converting heterogeneous data formats from the one or more external service platforms into a unified record format, deduplicating records based on identifier matching, and computing derived feature values comprising inter-event interval durations and cumulative quantity values.

### Claim 5 (depends on 4)

The system according to claim 4, wherein the circuitry is further configured to compute statistical summary values from the preprocessed records comprising mean interval, variance, trend direction, and seasonal pattern indicators for each monitored item.

### Claim 6 (depends on 1)

The system according to claim 1, wherein the comparison algorithm comprises normalizing the current condition data from the plurality of external information providing apparatuses to a common scale, computing a weighted score for each candidate based on a plurality of condition attributes, and ranking the candidates by the weighted score.

### Claim 7 (depends on 6)

The system according to claim 6, wherein the condition attributes comprise at least a primary value parameter, a discount parameter, a delivery time parameter, and a rating parameter, and the weights are derived from user preference data stored in the information storage device.

### Claim 8 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to acquire affective state data from the terminal device, input the affective state data into an emotion identification model to compute an emotion category and an emotion intensity value, and adjust at least one of the detail level or the length constraints in the prompt data structure based on the emotion category.

### Claim 9 (depends on 8)

The system according to claim 8, wherein the affective state data comprises at least one of facial image data, voice spectrogram data, or text sentiment data, and the emotion identification model comprises a neural network classifier that outputs the emotion category from a predefined set of emotion labels.

### Claim 10 (depends on 8)

The system according to claim 8, wherein the circuitry adjusts the prompt data structure by specifying a concise notification style when the emotion category indicates a busy or stressed state, and specifying a detailed explanation style when the emotion category indicates an engaged or relaxed state.

### Claim 11 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to acquire response data from the terminal device indicating user acceptance or rejection of the recommendation data, and generate relearning data from the response data for updating the time-series machine learning model.

### Claim 12 (depends on 11)

The system according to claim 11, wherein the circuitry is further configured to update the stored preference parameters based on the response data by increasing weight values for condition attributes associated with accepted recommendations and decreasing weight values for condition attributes associated with rejected recommendations.

### Claim 13 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to monitor the current condition data on a periodic schedule, detect a condition event when the effective value index for a monitored item satisfies a threshold condition, and trigger generation of the notification text data in response to the condition event.

### Claim 14 (depends on 13)

The system according to claim 13, wherein the circuitry is further configured to compute an urgency score for the condition event based on a proximity of the predicted time interval to a current timestamp and a magnitude of the effective value index, and prioritize notification generation for condition events having higher urgency scores.

### Claim 15 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to generate a summary report comprising aggregated historical transaction data, prediction accuracy metrics, and recommendation acceptance rates for transmission to the terminal device.

### Claim 16 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to associate the user identification data with a messaging service account via an authentication protocol, and transmit the generated notification text data to the terminal device through the messaging service account.

### Claim 17 (depends on 1)

The system according to claim 1, wherein the monitored items comprise household consumable products, the historical transaction records comprise purchase records from electronic commerce platforms, the current condition data comprises product price and availability data, and the generated notification text data comprises a purchase recommendation message.

### Claim 18

A system comprising:
circuitry configured to:
acquire user identification data and target item registration data from a terminal device via a communication interface, and store associated data in an information storage device;
collect historical transaction records from external service platforms, preprocess the records, and train a time-series machine learning model to generate prediction data for each monitored item;
acquire current condition data from a plurality of external information providing apparatuses, compute an effective value index, and determine recommendation data based on the prediction data;
construct a prompt data structure comprising the prediction data and the recommendation data, transmit the prompt data structure to a generative neural network model, and acquire generated notification text data; and
transmit the generated notification text data to the terminal device.

### Claim 19 (depends on 18)

The system according to claim 18, wherein the circuitry is further configured to acquire affective state data from the terminal device, compute an emotion category using an emotion identification model, and adjust the prompt data structure based on the emotion category.

### Claim 20

A method performed by circuitry via a communication interface coupled to a packet-switched network, the method comprising:
acquiring user identification data from a terminal device, associating the user identification data with account identification data of one or more external service platforms, and storing the associated identification data in an information storage device;
acquiring target item registration data comprising attribute data and periodicity data for monitored items;
collecting historical transaction records from the external service platforms, executing preprocessing comprising missing-value completion and timestamp normalization, and training a time-series machine learning model using the preprocessed records to generate prediction data;
acquiring current condition data from a plurality of external information providing apparatuses, computing an effective value index by applying a comparison algorithm, and determining recommendation data based on the prediction data;
constructing a prompt data structure by embedding the prediction data and the recommendation data into an instruction sequence, and transmitting the prompt data structure to a generative neural network model comprising a transformer-based architecture to acquire generated notification text data; and
transmitting the generated notification text data to the terminal device.
