17456168. IDENTIFYING PERSISTENT ANOMALIES FOR FAILURE PREDICTION simplified abstract (International Business Machines Corporation)

IDENTIFYING PERSISTENT ANOMALIES FOR FAILURE PREDICTION

Organization Name

International Business Machines Corporation

Inventor(s)

Seema Nagar of Bangalore (IN)

Pooja Aggarwal of Bengaluru (IN)

Dipanwita Guhathakurta of Kolkata (IN)

Rohan R. Arora of DANBURY CT (US)

Amitkumar Manoharrao Paradkar of Mohegan Lake NY (US)

Larisa Shwartz of Greenwich CT (US)

Bing Zhou of Rye NY (US)

Noah Zheutlin of White Plains NY (US)

IDENTIFYING PERSISTENT ANOMALIES FOR FAILURE PREDICTION - A simplified explanation of the abstract

This abstract first appeared for US patent application 17456168 titled 'IDENTIFYING PERSISTENT ANOMALIES FOR FAILURE PREDICTION

Simplified Explanation

The patent application describes a computer-based method and system for identifying persistent anomalies for failure prediction. The system receives a time series data stream and segments it into consecutive sliding windows. It then performs supervised and unsupervised persistent anomaly detection to determine if anomalies across the sliding windows are persistent. The results of both detection methods are combined to identify persistent anomalies.

• The system receives a time series data stream.
• It segments the data stream into consecutive sliding windows.
• It performs supervised persistent anomaly detection using a binary classification model.
• It performs unsupervised persistent anomaly detection.
• The results of both detection methods are combined to identify persistent anomalies.

Potential Applications

This technology can have various applications in different industries, including:

• Predictive maintenance in manufacturing: Identifying persistent anomalies can help predict failures in machinery and equipment, allowing for proactive maintenance and reducing downtime.
• Network monitoring and cybersecurity: Detecting persistent anomalies in network traffic can help identify potential security breaches or abnormal behavior.
• Financial fraud detection: Identifying persistent anomalies in financial transactions can help detect fraudulent activities and prevent financial losses.
• Healthcare monitoring: Monitoring patient data for persistent anomalies can help identify early signs of diseases or abnormalities.

Problems Solved

The technology addresses the following problems:

• Early failure prediction: By identifying persistent anomalies, the system can predict failures in advance, allowing for timely maintenance or intervention.
• Efficient anomaly detection: The combination of supervised and unsupervised detection methods improves the accuracy and efficiency of identifying persistent anomalies.
• Proactive decision-making: The system enables proactive decision-making by providing insights into potential failures or abnormal behavior.

Benefits

The technology offers several benefits:

• Improved reliability: By predicting failures in advance, the system helps improve the reliability and availability of machinery, systems, and networks.
• Cost savings: Proactive maintenance based on the identification of persistent anomalies can reduce downtime and repair costs.
• Enhanced security: Detecting persistent anomalies in network traffic can help prevent security breaches and protect sensitive data.
• Early detection of abnormalities: Identifying persistent anomalies in healthcare data can enable early detection and intervention for better patient outcomes.

Original Abstract Submitted

A computer-implemented method and a computer system for identifying persistent anomalies for failure prediction. The computer system receives a time series data stream. The computer system received a predetermined number N and a predetermined number M which is a fraction of N. The computer system segments the time series data stream into N consecutive sliding windows. The computer system performs supervised persistent anomaly detection to determine whether anomalies across the N consecutive sliding windows are persistent, by using a binary classification model. The computer system performs unsupervised persistent anomaly detection to determine whether the anomalies across the N consecutive sliding windows are persistent. The computer system combines results of the supervised persistent anomaly detection and results of the unsupervised persistent anomaly detection to determine persistent anomalies.