The development of dynamic information analysis methods, like incremental classification/clustering, concept drift management and novelty detection techniques, is becoming a central concern in a bunch of applications whose main goal is to deal with information which is varying over time or with information flows that can oversize memory storage or computation capacity. These applications relate themselves to very various and highly strategic domains, including web mining, social network analysis, adaptive information retrieval, anomaly or intrusion detection, process control and management recommender systems, technological and scientific survey, and even genomic information analysis, in bioinformatics.
The term “incremental” is often associated to the terms evolutionary, adaptive, interactive, on-line, or batch. The majority of the learning methods were initially defined in a non-incremental way. However, in each of these families, were initiated incremental methods making it possible to take into account the temporal component of a data flow or to achieve learning on huge/fast datasets in a tractable way. In a more general way incremental classification/clustering algorithms and novelty detection approaches are subjected to the following constraints:
- Potential changes in the data description space must be taken into consideration;
- Possibility to be applied without knowing as a preliminary all the data to be analyzed;
- Taking into account of a new data must be carried out without making intensive use of the already considered data;
- Result must but available after insertion of all new data.
The above mentioned constraints clearly follow the VVV (Volume-Velocity and Variety) rule and thus directly fit with big/fast data context.
This workshop aims to offer a meeting opportunity for academics and industry-related researchers, belonging to the various communities of Computational Intelligence, Machine Learning, Experimental Design, Data Mining and Big/Fast Data Management to discuss new areas of incremental classification, concept drift management and novelty detection and on their application to analysis of time varying information and huge dataset of various natures. Another important aim of the workshop is to bridge the gap between data acquisition or experimentation and model building.
Through an exhaustive coverage of the incremental learning area workshop will provide fruitful exchanges between plenaries, contributors and workshop attendees. The emerging big/fast data context will be taken into consideration in the workshop.
The set of proposed incremental techniques includes, but is not limited to:
- Novelty detection algorithms and techniques
- Semi-supervised and active learning approaches
- Machine learning for data streams
- Adaptive hierarchical, k-means or density-based methods
- Adaptive neural methods and associated Hebbian learning techniques
- Incremental deep learning
- Multiview diachronic approaches
- Probabilistic approaches
- Distributed approaches
- Graph partitioning methods and incremental clustering approaches based on attributed graphs
- Incremental clustering approaches based on swarm intelligence and genetic algorithms
- Evolving classifier ensemble techniques
- Incremental classification methods and incremental classifier evaluation
- Dynamic feature selection techniques
- Clustering of time series
- Visualization methods for evolving data analysis results
The list of application domain is includes, but it is not limited to:
- Evolving textual information analysis
- Evolving social network analysis
- Dynamic process control and tracking
- Intrusion and anomaly detection
- Genomics and DNA microarray data analysis
- Adaptive recommender and filtering systems
- Scientometrics, webometrics and technological survey