18125268. AUTOMATED FEATURE ENGINEERING simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)

AUTOMATED FEATURE ENGINEERING

Organization Name

INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor(s)

Kunal Sawarkar of Franklin Park NJ (US)

Shivam Raj Solanki of Austin TX (US)

Christopher Chen of San Jose CA (US)

Amit P. Joglekar of Olathe KS (US)

AUTOMATED FEATURE ENGINEERING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18125268 titled 'AUTOMATED FEATURE ENGINEERING

• • Simplified Explanation:**

The patent application discusses feature engineering in automated machine learning, focusing on processing streaming data from sensors to detect anomalies.

• • Key Features and Innovation:**
• Receive streaming data from sensors representing attributes over time.
• Compute long term and short term point statistics from the data.
• Quantize the data into time windows and compute statistics based on intervals.
• Normalize the long term and short term statistics.
• Apply dynamic time warping to compare the normalized statistics.
• Generate probability distributions based on the comparison.
• Produce machine learning input features based on the distance between the mean values of the distributions.
• Train a machine learning model to detect anomalies using the input features.

• • Potential Applications:**

This technology can be applied in various industries such as manufacturing, healthcare, finance, and cybersecurity for anomaly detection in real-time data streams.

• • Problems Solved:**

The technology addresses the challenge of efficiently processing streaming data from sensors and detecting anomalies in a timely manner.

• • Benefits:**
• Improved accuracy in anomaly detection.
• Real-time monitoring and alerting for potential issues.
• Enhanced efficiency in processing large volumes of streaming data.

• • Commercial Applications:**

Potential commercial applications include predictive maintenance in manufacturing, fraud detection in finance, and health monitoring in healthcare industries.

• • Prior Art:**

Researchers can explore prior art related to anomaly detection in streaming data processing, dynamic time warping, and machine learning input feature generation.

• • Frequently Updated Research:**

Stay updated on advancements in anomaly detection algorithms, sensor technology, and machine learning models for real-time data processing.

• • Questions about Feature Engineering in Automated Machine Learning:**

1. How does dynamic time warping improve anomaly detection in streaming data processing? 2. What are the key challenges in normalizing long term and short term point statistics for feature engineering in automated machine learning?

Original Abstract Submitted

Feature engineering, for example, in automated machine learning, can include receiving streaming data representing at least one attribute detected by a sensor over time. Long term point statistics associated with the streaming data can be computed. The streaming data can be quantized into intervals of time windows and short term point statistics based on the intervals can be computed. The long term point statistics and the short term point statistics can be normalized. Dynamic time warping can be applied across the normalized long term point statistics and short term point statistics. A pair of probability distributions can be generated associated with the dynamic time warped normalized long term point statistics and short term point statistics. Based on distance between the mean values of the probability distributions, machine learning input features can be produced. The machine learning input features can be fed to train a machine learning model for detecting anomaly.