Associate Professor
Information Extraction from Free Text Data in Health
What You'll Learn
- Identify text mining approaches needed to identify and extract different kinds of information from health-related text data.
- Differentiate how training deep learning models differ from training traditional machine learning models.
4 Modules
24 Hours
6 hrs per module (approx.)
Rating
About Information Extraction from Free Text Data in Health
In this MOOC, you will be introduced to advanced machine learning and natural language
processing techniques to parse and extract information from unstructured text documents in
healthcare, such as clinical notes, radiology reports, and discharge summaries. Whether you are an aspiring data scientist or an early or mid-career professional in data science or information technology in healthcare, it is critical that you keep up-to-date your skills in information extraction and analysis.
To be successful in this course, you should build on the concepts learned through other intermediate-level MOOC courses and specializations in Data Science offered by the University of Michigan, so you will be able to delve deeper into challenges in recognizing medical entities in health-related documents, extracting clinical information, addressing ambiguity and polysemy to tag them with correct concept types, and develop tools and techniques to analyze new genres of health information.
By the end of this course, you will be able to:
Identify text mining approaches needed to identify and extract different kinds of information from health-related text data
Create an end-to-end NLP pipeline to extract medical concepts from clinical free text using one terminology resource
Differentiate how training deep learning models differ from training traditional machine learning models
Configure a deep neural network model to detect adverse events from drug reviews
List the pros and cons of Deep Learning approaches."
Skills You'll Gain
- Data Mining
- Information Extraction
- Machine Learning
- Python (Programming Language)
What You'll Earn
- Certificate of Completion:
- Certificates of completion acknowledge knowledge acquired upon completion of a non-credit course or program.
Welcome Message
Information Extraction from Free Text Data in Health introduces machine learning and natural language processing techniques for analyzing unstructured healthcare data. Learners develop hands-on skills to extract clinical information from medical text while addressing ambiguity, privacy, and real-world healthcare data challenges.
This abbreviated syllabus description was created with the help of AI tools and reviewed by staff. The full syllabus is available to those who enroll in the course.
Course Schedule
Module 1: What is Information Extraction
- Video: Welcome to Information Extraction from Free Text Data in Health
- Reading: Syllabus
- Discussion Prompt: Meet your Classmates
- Reading: Community Engagement Rules
- Reading: Help Us Learn More About You
- Video: What is Information Extraction? | Part 1
- Video: What is Information Extraction? | Part 2
- Video: Information Extraction on Formatted Text
- Discussion Prompt: Exercise 1: Variety of Date Formats
- Video: Identifying Dates
- Video: Using Curated Lists for Information Extraction
- Video: Evaluation Metrics
- Discussion Prompt: Exercise 2: Power of Lists
- Video: Hands-On Exercise Demo
Module 2: Named Entity Recognition (NER)
- Video: Medical Natural Language Processing | Part 1
- Video: Medical Natural Language Processing | Part 2
- Discussion Prompt: Applications of Language Processing Steps in Medicine
- Video: Health Ontology Resources | Part 1
- Video: Health Ontology Resources | Part 2
- Video: Health Ontology Resources | Part 3
- Discussion Prompt: Health Ontology Resources: Building a Concept Extraction Pipeline
- Video: Hands-On Exercise Demo
Module 3: Sequential Classification
- Video: Introduction to Medical Named Entity Extraction
- Video: Medical Named Entity Extraction
- Discussion Prompt: Building De-Identification Toolkit
- Video: Sequence Labeling
- Video: Hidden Markov Models
- Graded Assignment: Hidden Markov Models: Knowledge Check
- Discussion Prompt: Hidden Markov Models and Selected Applications in Speech Recognition
- Video: Conditional Random Fields
- Video: NER Features
- Graded Assignment: Designing Features for a Conditional Random Fields Model: Knowledge Check
- Video: Hands-On Exercise Demo
Module 4: Introduction to Advanced Approaches to NER in Health
- Video: What is Deep Learning?
- Video: Perceptron: Simplest Neural Network
- Graded Assignment: Perceptron: Knowledge Check
- Video: Deep Neural Networks
- Video: Deep Learning: Applications
- Graded Assignment: Deep Neural Network Models: Knowledge Check
- Video: Hands-On Exercise Demo
- Reading: Post-Course Survey
Grading Policy
Learners must earn at least 80% overall to pass. Four hands-on programming exercises account for 95% of the grade, with a module 1 quiz worth 5% completing the total course grade.
Course content developed by U-M faculty and managed by the university. Faculty titles and affiliations are updated periodically.
Intermediate Level
Some related experience required
