Friday, 16:15 - 16:40 h, Room: MA 376


Giovanni Felici
Logic data mining in the presence of noisy data

Coauthor: Emanuel Weitschek


In this work we consider a method for the extraction of knowledge from data. The knowledge is represented as disjunctive normal form (DNF) logic formulas that identify with high precision subsets of the training data. The method is mainly designed for classification purposes, but can be profitably deployed for information compression and data analysis in general. It is based on three main steps: discretization, feature selection and formula extraction. For each step, a mathematical optimization problem is formulated and solved with ad hoc algorithmic strategies.
The method is designed to perform exact separation of training data, and can thus be exposed to overfitting when a significant amount of noise is present in the available information. We analyze the main problems that arise when this method deals with noisy data and propose extensions to the discretization, feature selection and formula extraction steps; we motivate these extensions from a theoretical standpoint, and show with experimental evidence how they operate to remove the effect of noise on the mining process.


Talk 3 of the invited session Fri.3.MA 376
"Methods from discrete mathematics in systems biology" [...]
Cluster 12
"Life sciences & healthcare" [...]


  Payday Loans In Ohio. Since its introduction in the market buying Cialis can be exclusively in pharmacy chains with a prescription from a doctor. I agree that this is very inconvenient and takes a lot of time and effort.