Record Detail Back
Feature Selection for High-Dimensional Data
- Share to:
The topic of variable selection in high-dimensional spaces (often with hundreds or thousands of dimensions) has attracted considerable attention in data mining re- search in previous years, and it is common in many real problems.
In a nutshell, feature selection is a process that chooses an optimal subset of features according to a certain criterion. The selection of the criterion must be done according to the purpose of feature selection, usually with the aim of improving the prediction accuracy of the data mining algorithm used to learn a model. Generally, the objective is to identify the features in the dataset which are important and discard others as redundant or irrelevant. The problem is especially relevant when we are managing a huge number of features and the learning algorithm loses prediction capacity using all of them. Since feature selection reduces the dimensionality of the data, the data mining algorithms can run faster and obtain better outcomes by using feature selection.
The publication of the book “Feature Selection for High-Dimensional Data” written by Vero ́nica Bolo ́n-Canedo, Noelia Sa ́nchez-Maron ̃o and Amparo Alonso- Betanzos is an important event. The book offers a coherent and comprehensive ap- proach to feature subset selection in the scope of classification problems.
We can shortly outline the three parts found when reading the book: foundations, real application problems and challenges. First, the authors focus on the analysis and synthesis of feature selection algorithms, presenting a comprehensive review of basic concepts and experimental results of the most well known algorithms. Sec- ond, an interesting novelty and contribution of the book is how it addresses different real scenarios with high-dimensional data, showing the use of feature selection algo- rithms in different contexts with different requirements and information: microarray data, intrusion detection, tear film lipid layer classification and cost based features. Third, the book also delves into the scenario of big dimension, sometimes com- bined with massive amounts of data as big data. It pays attention to important prob- lems under high-dimensional spaces: scalability, distributed processing and real- time processi
1st Edtion
978-3-319-21858-8
NONE
Feature Selection for High-Dimensional Data
Information Technology
English
Springer International Publishing
2015
Feature Selection for High-Dim
1-168
LOADING LIST...
LOADING LIST...