Our seminars take place on Wednesdays between 12:00 - 2:00 PM in the Center for Innovation and Transfer of Natural Sciences and Engineering Knowledge (building A0, wing B1, room 228).

Sign up to our newsletter

For your convenience we would like to kindly invite you to sign up to our newsletter. After filling in the fields below and submitting the form, please check your e-mail account and confirm the registration.
Email *

Visit counter

Online: 1
Total: 34746

May 10, 2017, Dr. Wiesław Paja, “Generational feature elimination to find all relevant feature subset”

wpAbstract: The recent increase of dimensionality of data is a target for many existing feature selection methods with respect to efficiency and effectiveness. In this pesentation, the all relevant feature selection method based on information gathered using generational feature elimination will be introduced. The successive generations of feature subset were defined using DTLevelImp algorithm and in each step the subset of most important features were eliminated from the primary investigated dataset. This process was executed until the most important feature reach importance value on the level similar to importance of the random shadow features. The proposed method was also initially tested on well-know artificial and real-world datasets and the results confirm its efficiency. Thus, it can be concluded that selected attributes are relevant.

Comments are closed.