Call for Papers: Special Issue on Feature Engineering
The main objective of machine learning is to extract patterns to turn data into knowledge. Since the beginning of this century, technological advances have drastically changed the size of data sets as well as the speed with which these data must be analyzed. Modern data sets may have a huge number of instances, a very large number of features, or both. In most applications, data sets are compiled by combining data from different sources and databases (containing both structured and unstructured data) where each source of information has its strengths and weaknesses. Before applying any machine learning algorithm, it is therefore necessary to create interesting features from the data sources. This essential step, which is denoted “feature engineering” or feature extraction, is of utmost importance in the machine learning process. Machine learners should be well aware of the power of feature engineering and it is important to share good practices.
This special issue aims to bring together innovative feature engineering techniques and/or successful development of features that improve the performance and/or interpretability of machine learning models. Both manual (relying on human creativity and/or domain knowledge) as well as automated (obtained for example from a relational dataset) feature engineering techniques are considered. We encourage novel featurization techniques that are based on diverse and alternative data sources and that leverage the temporal, granular, or unstructured aspects of the data.
Topics of Interest
We welcome original research papers on all aspects of feature engineering including, but not limited to the following topics:
Extracting meaningful features from transactional data, network data, textual data, unstructured data, and/or temporal data
Constructing features using domain knowledge
Automatic feature creation
Feature engineering for anomaly detection
Feature engineering to reduce/avoid bias
Risks of feature engineering
Information/performance gain using correct feature engineering (case studies)
Contributions must contain new, unpublished, original and fundamental work relating to the Machine Learning Journal’s mission. Purely theoretical papers without thorough empirical evaluation, simple surveys, and/or incremental contributions are discouraged. All submissions will be reviewed using rigorous scientific criteria whereby the novelty of the contribution will be crucial.