Transactional data is ubiquituous in many real world applications. While market-basket, protein-taxa, gene expression, etc. naturally present themselves as transactional data, others like streaming/time-series data, text data, protein-protein interactions, etc. can also be modeled as transactional data. Transactional databases are typically large and have high dimensionality, and may also be affected by noise. Noise may arise due to missing values, erroneous readings or uncertainty in data. Depending on the domain, the noise may be either known or unknown (both in distribution and in amount).
In this workshop, we focus on common mining problems: Frequent itemset mining, Association Rule Mining, Subspace clustering and Co-clustering, on transactional databases in the presence of noise. In such problems, noise can lead to incorrect mining results and also degrade mining performance. So it is critical to develop noise aware mining techniques which are robust and efficient. Besides developing innovative techniques, it is also important to validate them using statistical/domain knowledge or by real-world case studies.