Search

• Propose "bidirectional" analysis and novelty validation standard based on Gap statistic and Isolation forest. • The number of unknown targets (especially high-correlated) in open world could be ascertained accurately and flexibly. • Some advantages, e.g., no multiple choice issue, data-driven, objective observation, wide applicability could be shown. How to ascertain the number of unknown targets in open world is a crucial issue in generalized evidence theory (GET). To this, limited works have been done, and they may trap in low accuracy, multi-choice issue, complex implementation, etc. To overcome the above limitations, a new method which introduces Isolation forest and Gap statistic into K-means method to perform "bidirectional" analysis is proposed in this paper. According to the "forward" clustering result, isolation-based anomaly detection could achieve the partition of normal domain and abnormal domain, then "reverse" clustering could directly focus on the abnormal domain to pick out those novel clusters using the newly defined novelty validation standard. Further, the defective frame of discernment (FOD) could be completed since the sum of base clusters and novel clusters is the actual target numbers. The proposed method could guarantee the integrity of FOD even the features of the samples are highly-correlated. Various simulation experiments illustrate its effectiveness and wide applicability. [ABSTRACT FROM AUTHOR]

Abstract: We address the problem of designing quantizers that cluster data while satisfying a k-anonymity requirement. A general data compression perspective is adopted, which considers both discrete and continuous probability distributions, and corresponding constraints on both cell sizes and quantizer index probabilities. Potential applications of this problem extend well beyond the important case of microdata anonymization, to include also optimized task allocation under workload constraints. Our contribution is twofold. First and most importantly, we present a theoretical analysis showing the optimality conditions which probability-constrained quantizers must satisfy, thereby theoretically characterizing optimal k-anonymous aggregation as a special case. As a second contribution, inspired by our theoretical analysis, we propose an alternating optimization algorithm for the design of this type of quantizers. Our algorithm is conceptually motivated by the popular Lloyd–Max algorithm for quantization design, originally intended for data compression, also known as the k-means method. Experimental results for synthetic and real data, with mean squared error as a distortion measure, confirm that our method outperforms MDAV, a popular fixed-size microaggregation algorithm for statistical disclosure control. This performance improvement is in terms of data utility, for the exact same k-anonymity constraint, but does come at the expense of higher computational sophistication. [Copyright &y& Elsevier]

With the exponential growth of information on the World Wide Web, there is great demand for developing efficient methods for effectively organizing the large amount of retrieved information. Document clustering plays an important role in information retrieval and taxonomy management for the Web. In this paper we examine three clustering methods: K-means, multi-level METIS, and the recently developed normalized-cut method using a new approach of combining textual information, hyperlink structure and co-citation relations into a single similarity metric. We found the normalized-cut method with the new similarity metric is particularly effective, as demonstrated on three datasets of web query results. We also explore some theoretical connections between the normalized-cut method and the K-means method. [Copyright &y& Elsevier]