WEBVTT

00:01.320 --> 00:02.240
Hello everyone!

00:02.840 --> 00:08.720
In this video we will learn about Analyzer Engine and take a deep dive on it.

00:09.880 --> 00:17.120
The Procedures Analyzer engine is the intelligent hub for PII detection.

00:17.960 --> 00:27.240
It leverages sophisticated combination of techniques to accurately locate sensitive entities, even

00:27.560 --> 00:30.040
in a complex and unstructured data.

00:31.680 --> 00:38.720
The very first technique that we'll talk about is pattern matching.

00:39.800 --> 00:46.680
It uses regular expressions for structured entities like credit card numbers, phone numbers, and Social

00:46.680 --> 00:48.840
Security numbers.

00:51.400 --> 01:00.760
The second detection technique is NLP models, where it uses named entity recognition for context aware

01:00.800 --> 01:13.320
detections of names and locations it Also uses deny lists for custom word list for organization specific

01:13.320 --> 01:22.840
sensitive terms, and then it also uses context analysis using surrounding words and phrases to increase

01:22.840 --> 01:27.240
detection accuracy and reduce false positives.

01:29.600 --> 01:38.440
Each of these techniques returns the entity type, location in the text and a confidence score that

01:38.440 --> 01:40.200
is between 0 and 1.

01:42.400 --> 01:44.880
So what about the extensibility?

01:45.320 --> 01:55.160
The built in custom recognizers for domain specific PII is extensible to internal employee IDs, customer

01:55.160 --> 02:00.480
account numbers, medical records, and proprietary data formats.

02:01.800 --> 02:07.800
Custom recognizers integrate seamlessly with existing data detection pipelines.

02:08.760 --> 02:09.440
Thank you.

02:09.480 --> 02:10.880
I'll see you in the next video.