last words dm 1. mining data steams / incremental data mining / mining sensor data (e.g. modify a...

4
Last Words DM 1. Mining Data Steams / Incremental Data Mining / Mining sensor data (e.g. modify a decision tree assuming that new examples arrive continuously, and old examples are discarded) 2. Text Mining 3. Mining the Web/Mining Graphs and other complex structures 4. Mining spatial-temporal data, particularly environmental, cell-phone, and traffic data 5. Contrast mining (e.g. how do two groups of people differ) 6. Data Mining and Privacy 7. Mining Social Networks (kind of hot these days) 8. Statistical Techniques (Principal component analysis, multi-dimensional scaling, feature selection, statistical testing, Bayesian classifier,...)typically taught in a Machine Learning class. 9. Preprocessing probably deserves more coverage 10. High Performance Data Mining Parallel Programming Course Mining that we didn’t or very little discuss in this class

Upload: tamsyn-page

Post on 04-Jan-2016

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Last Words DM 1. Mining Data Steams / Incremental Data Mining / Mining sensor data (e.g. modify a decision tree assuming that new examples arrive continuously,

Last Words DM

1. Mining Data Steams / Incremental Data Mining / Mining sensor data (e.g. modify a decision tree assuming that new examples arrive continuously, and old examples are discarded)

2. Text Mining 3. Mining the Web/Mining Graphs and other complex structures 4. Mining spatial-temporal data, particularly environmental, cell-phone,

and traffic data 5. Contrast mining (e.g. how do two groups of people differ)6. Data Mining and Privacy 7. Mining Social Networks (kind of hot these days) 8. Statistical Techniques (Principal component analysis, multi-

dimensional scaling, feature selection, statistical testing, Bayesian classifier,...)typically taught in a Machine Learning class.

9. Preprocessing probably deserves more coverage10. High Performance Data Mining Parallel Programming Course

Other Important Topics in Data Mining that we didn’t or very little discuss in this class

Other Important Topics in Data Mining that we didn’t or very little discuss in this class

Page 2: Last Words DM 1. Mining Data Steams / Incremental Data Mining / Mining sensor data (e.g. modify a decision tree assuming that new examples arrive continuously,

Last Words DM

1. Develop a unifying theory for data mining (e.g. explaining how and when over-fitting occurs)

2. Mining data streams / mining sensor networks / mining sequential data3. High performance data mining platforms / combining parallel

computing and data mining (http://en.wikipedia.org/wiki/Hadoop) 4. Spatial data mining / temporal data mining / spatial temporal 5. Mining graphs and other complex types of data6. More research on the interestingness of knowledge7. Distributed data mining (cannot pass the complete data set; distributed

decision making, e.g. in sensor networks)8. Data mining for genomic and earth science problems9. What is the data mining process --- kind of software engineering for

data mining; development of data mining methodologies…10. Data Mining without violating privacy and security

New Challenges for the Field of Data MiningNew Challenges for the Field of Data Mining

Page 3: Last Words DM 1. Mining Data Steams / Incremental Data Mining / Mining sensor data (e.g. modify a decision tree assuming that new examples arrive continuously,

Last Words DM

Complementary Knowledge For Getting Jobs in Data Mining

Data Mining

Databases

Data Structures

& Algorithms

Software Design

Machine Learning

AI

High Performance

Computing

Evolutionary

Computing

Pattern Recognition

Statistics

Optimization

Information Retrieval

Image Processing

GIS

Data Visualization

Search Techniques

Experimental

Evaluation

Software Engineering

Page 4: Last Words DM 1. Mining Data Steams / Incremental Data Mining / Mining sensor data (e.g. modify a decision tree assuming that new examples arrive continuously,

Last Words DM

2008 Student Textbook Evaluation Overall positive evaluation but

– Some felt that algorithms were not explained in sufficient detail, particularly examples are missing

– A few felt the material should be better indexed

– Some felt it lack highlighting of key points

– Some felt it is at an intermediate level, and does not give sufficient depth if the textbook is your only source of knowledge; it also introduces topics more intuitively and not formally, as some more advanced textbook do.

2 students felt that the textbook does not introduce topics very clearly, and that it is not comprehensive.