2016/10/06 by Muhammad Imran, Imran, Muhammad, Sanjay Chawla +2
Computer Science · Engineering · #Anomaly Detection Techniques and Applications #Data Stream Mining Techniques #Artificial Immune Systems Applications
paper · pdf · doi:10.48550/arxiv.1610.01858
An emerging challenge in the online classification of social media data\nstreams is to keep the categories used for classification up-to-date. In this\npaper, we propose an innovative framework based on an Expert-Machine-Crowd\n(EMC) triad to help categorize items by continuously identifying novel concepts\nin heterogeneous data streams often riddled with outliers. We unify constrained\nclustering and outlier detection by formulating a novel optimization problem:\nCOD-Means. We design an algorithm to solve the COD-Means problem and show that\nCOD-Means will not only help detect novel categories but also seamlessly\ndiscover human annotation errors and improve the overall quality of the\ncategorization process. Experiments on diverse real data sets demonstrate that\nour approach is both effective and efficient.\n