Syllabus - 80

advertisement
The Chaim Sheba Medical Center
Tel Hashomer
ISRAEL
The Biological Informatics Program
The Mina & Everard Goodman
Faculty of Life Sciences
Bar-Ilan University
ISRAEL
Medical Data Mining - Course Number 80665
Extent: Lecture - 2 weekly hours; Exercise - 1 weekly hour
Updated: January, 2012
Course Description
The use of algorithms to mine medical data, information and knowledge, as it is
diverse and scattered among EMRs, medical journals, social networks and the web in
general, has long been one of the most desired goals of Artificial Intelligence (A.I.).
The potential applications in real-life situations and the implications to the medical
industry, if and when algorithms function as good as we intend, are enormous and
can hardly be grasped or imagined.
Moreover, thanks to the growing spread of A.I. algorithms throughout the web, and
its close and quick connection to large revenues (such as A.I. for online advertising,
ranking search results, recommendation engines in e-commerce sites, etc.), data
mining algorithms have recently evolved, and in a tremendous scale.
The course is built as a dialog between horizontal and vertical subjects, being the
medical applications of data mining algorithms, on one hand, and the computational
methodologies and fundamental theories that are intertwined in each of those
applications, on the other hand.
Detailed Schedule
1] Vertical subjects - medical applications of data mining algorithms:
Week
Topic
1
Introduction I
2
Introduction II
3
Data  Information 
Knowledge
4
EMRs I
5
EMRs II
6
Web & Smartphones I
7
Web & Smartphones II
8
Telemedicine
9
Artificial Intelligence I
10
Artificial Intelligence II
11
Artificial Intelligence III
12
Personalized Medicine I
13
Personalized Medicine II
Details
Medical Data Mining as a discipline, history,
major players, legal aspects.
Obstacles and challenges, current state of the
art in research institutes, hospitals, clinics, the
business sector and patients’ personal
computers.
Types, scope, gathering techniques, filtering,
interpretation, intelligent analysis and
knowledge management.
Motivation, history, current state of the art,
the governmental perspective, business side,
insurance issues.
Data standardization, “client hierarchies”:
patients, physicians, medical centers and
research institutes, as well as the role of
private start-ups.
Institutional medical information, patients’
forums, social networks and wisdom of the
crowd.
Mobile measurement techniques, wearable
elements, web crawling, symptom analyzers
and recommendation systems.
Patient-patient, patient-physician and
physician-physician communication, education,
robotics, virtual reality and remote surgery.
Motivation, evidence-based medicine,
limitation of human analysis and inference,
mistake detection technologies.
Diagnosis, expert systems, clinical decision
support, hypothesis engines and
personalization algorithms.
Machine learning, text analysis and mining,
natural language processing, automatic sideeffect identification and social data mining.
History, Medical Informatics as the basis of
personalized medicine, role of Bioinformatics,
prediction of treatment outcome.
Mathematical models, response-based
medicine and current state of the art at clinical
trials and the private sector.
2] Horizontal subjects - computational methodologies and fundamental theories:
Week
1
2
3
4
5
6
7
8
9
10
11
12
13
Topic
Data representation options
SQL – Standard Query Language
Information retrieval (IR) engines
Pattern recognition and motif discovery
Clustering and classifiers
Weka 3
Machine learning: artificial neural networks
Machine learning: support vector machines
Multi-dimensional search
Principal Component Analysis
Basic graph theory
Decision trees
‘ePocrates’, ‘AOTrauma’, ‘Patients Like Me’, ‘23andMe and ‘Knome’