# US 2021/0080939 A1 · Anomaly detection using mset with random projections

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

- Number: US 2021/0080939 A1
- Document ID: US20210080939A1
- Status: Published application
- Kind: A1
- Application number: 16/948,337
- Filed: 2020-09-14
- Published: 2021-03-18
- Earliest claimed: 2019-09-15
- Assignees: Oracle International Corporation
- Applicants: Oracle International Corporation
- Inventors: Joe YARMUS; Boriana MILENOVA
- Web page: https://patentgrep.com/patent/US20210080939A1

## CPC

- G05B 23/0221: {Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods}
- G05B 23/0254: {based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks}
- G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks}
- G05B 23/0283: {Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] (preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06)}
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)

## Abstract

Disclosed is an approach to implement improved anomaly detection. Improved anomaly detection is provided using MSET-SPRT via Monte Carlo simulation that can address problems with conventional MSET-SPRT approaches and provide improved system performance and accuracy.

## Claims

### Claim 1

A method, comprising:
gathering sensor data;
consolidating a per-feature statistical analysis for the sensor data into one or more global or region aggregates;
using random projection to reduce dimensionality of the sensor data into the global or region aggregates and to interpret projections as generated Monte Carlo estimates of the aggregates; and
performing anomaly identification using a result of global or region hypothesis testing.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
using random projection to reduce the sensor data from individual data to the global or region aggregates by division of sensors into sensor regions; and
building an MSET (Multivariate State Estimation Technique) on the random projection.

### Claim 3 (depends on 1)

The method of claim 1, wherein the random projection corresponds to a fixed boundary, Monte Carlo simulation that generates valid p-values for a binomial test of global or regional health

### Claim 4 (depends on 1)

The method of claim 1, wherein the global or region hypothesis testing comprises a single hypothesis.

### Claim 5 (depends on 1)

The method of claim 1, wherein Monte Carlo simulation is used to reduce both false alerts and failures to detect anomalous behavior.

### Claim 6 (depends on 1)

The method of claim 1, wherein the sensor data comprises data from IoT sensors.

### Claim 7 (depends on 1)

The method of claim 1, wherein random projections is performed to reduce the dimensionality of the sensor data to generate reduced dimensionality data.

### Claim 8 (depends on 1)

The method of claim 1, wherein each random projection corresponds to a Monte Carlo simulation of system health.

### Claim 9 (depends on 1)

The method of claim 1, wherein individual sensor behaviors from multiple sensors are merged into a MSET model generated by reduced dimensionality data, and a single hypothesis is tested to identify a global anomaly across the multiple sensors.

### Claim 10 (depends on 1)

The method of claim 1, wherein the anomaly identification is performed by:
obtaining a prediction from MSET/SPRT;
counting violations; and
applying a binomial test.

### Claim 11 (depends on 1)

The method of claim 1, wherein a threshold number is established such that reduced dimensionality data is generated to perform the anomaly identification on a global basis when a number of sensors exceeds the threshold number, and full sensor data used to identify sensor-by-sensor anomalies when the number of sensors is lower than the threshold number.

### Claim 12

A system, comprising:
a processor;
a memory for holding programmable code; and
wherein the programmable code includes instructions executable by the processor for gathering sensor data; consolidating a per-feature statistical analysis for the sensor data into one or more global or region aggregates; using random projection to reduce dimensionality of the sensor data into the global or region aggregates and to interpret projections as generated Monte Carlo estimates of the aggregates; and performing anomaly identification using a result of global or region hypothesis testing.

### Claim 13 (depends on 12)

The system of claim 12, wherein the programmable code includes additional instructions for:
using random projection to reduce the sensor data from individual data to the global or region aggregates by division of sensors into sensor regions; and
building an MSET (Multivariate State Estimation Technique) on the random projection.

### Claim 14 (depends on 12)

The system of claim 12, wherein the random projection corresponds to a fixed boundary, Monte Carlo simulation that generates valid p-values for a binomial test of global or regional health

### Claim 15 (depends on 12)

The system of claim 12, wherein the global or region hypothesis testing comprises a single hypothesis.

### Claim 16 (depends on 12)

The system of claim 12, wherein Monte Carlo simulation is used to reduce both false alerts and failures to detect anomalous behavior.

### Claim 17 (depends on 12)

The system of claim 12, wherein the sensor data comprises data from IoT sensors.

### Claim 18 (depends on 12)

The system of claim 12, wherein random projections is performed to reduce the dimensionality of the sensor data to generate reduced dimensionality data.

### Claim 19 (depends on 12)

The system of claim 12, wherein each random projection corresponds to a Monte Carlo simulation of system health.

### Claim 20 (depends on 12)

The system of claim 12, wherein individual sensor behaviors from multiple sensors are merged into a MSET model generated by reduced dimensionality data, and a single hypothesis is tested to identify a global anomaly across the multiple sensors.

### Claim 21 (depends on 12)

The system of claim 12, wherein the programmable code includes additional instructions for performing anomaly detection by:
obtaining a prediction from MSET/SPRT;
counting violations; and
applying a binomial test.

### Claim 22 (depends on 12)

The system of claim 12, wherein a threshold number is established such that reduced dimensionality data is generated to perform the anomaly identification on a global basis when a number of sensors exceeds the threshold number, and full sensor data used to identify sensor-by-sensor anomalies when the number of sensors is lower than the threshold number.

### Claim 23

A computer program product embodied on a computer readable medium, the computer readable medium having stored thereon a sequence of instructions which, when executed by a processor, executes a method comprising:
gathering sensor data;
consolidating a per-feature statistical analysis for the sensor data into one or more global or region aggregates;
using random projection to reduce dimensionality of the sensor data into the global or region aggregates and to interpret projections as generated Monte Carlo estimates of the aggregates; and
performing anomaly identification using a result of global or region hypothesis testing.

### Claim 24 (depends on 23)

The computer program product of claim 23, wherein the sequence of instructions further causes:
using random projection to reduce the sensor data from individual data to the global or region aggregates by division of sensors into sensor regions; and
building an MSET (Multivariate State Estimation Technique) on the random projection.

### Claim 25 (depends on 23)

The computer program product of claim 23, wherein the random projection corresponds to a fixed boundary, Monte Carlo simulation that generates valid p-values for a binomial test of global or regional health

### Claim 26 (depends on 23)

The computer program product of claim 23, wherein the global or region hypothesis testing comprises a single hypothesis.

### Claim 27 (depends on 23)

The computer program product of claim 23, wherein the sequence of instructions further uses Monte Carlo simulation to reduce both false alerts and failures to detect anomalous behavior.

### Claim 28 (depends on 23)

The computer program product of claim 23, wherein the sensor data comprises data from IoT sensors.

### Claim 29 (depends on 23)

The computer program product of claim 23, wherein the sequence of instructions performs random projections to reduce the dimensionality of the sensor data to generate reduced dimensionality data.

### Claim 30 (depends on 23)

The computer program product of claim 23, wherein each random projection corresponds to a Monte Carlo simulation of system health.

### Claim 31 (depends on 23)

The computer program product of claim 23, wherein individual sensor behaviors from multiple sensors are merged into a MSET model generated by reduced dimensionality data, and a single hypothesis is tested to identify a global anomaly across the multiple sensors.

### Claim 32 (depends on 23)

The computer program product of claim 23, wherein the sequence of instructions further performs anomaly detection by:
obtaining a prediction from MSET/SPRT;
counting violations; and
applying a binomial test.

### Claim 33 (depends on 23)

The computer program product of claim 23, wherein the sequence of instructions further establishes a threshold number such that reduced dimensionality data is generated to perform the anomaly identification on a global basis when a number of sensors exceeds the threshold number, and full sensor data used to identify sensor-by-sensor anomalies when the number of sensors is lower than the threshold number.

### Claim 34

A method, comprising:
gathering sensor data;
reducing dimensionality of the sensor data to generate reduced dimensionality data;
constructing a MSET (Multivariate State Estimation Technique) model using the reduced dimensionality data;
performing hypothesis testing using a single hypothesis; and
performing anomaly identification using a result of the hypothesis testing.

### Claim 35 (depends on 34)

The method of claim 34, wherein random projections is performed to reduce the dimensionality of the sensor data to generate the reduced dimensionality data.

### Claim 36 (depends on 35)

The method of claim 35, wherein each random projection corresponds to a Monte Carlo simulation of system health.

### Claim 37 (depends on 34)

The method of claim 34, wherein individual sensor behaviors from multiple sensors are merged into the MSET model generated by the reduced dimensionality data, and the single hypothesis is tested to identify a global anomaly across the multiple sensors.
