Direct local pattern sampling by efficient two-step random procedures

Boley, Mario; Lucchese, Claudio; Paurat, Daniel; Gärtner, Thomas

doi:10.1145/2020408.2020500

Cited by 59 publications

(132 citation statements)

References 31 publications

Supporting

Mentioning

132

Contrasting

Order By: Relevance

“…In order to adopt this algorithm for a data stream, we use T ile to recalculate the top-k largest tiles from scratch whenever the window is sliding. -sTile: a sampling based technique proposed in [2]. The sT ile algorithm samples N itemsets from the window such that each itemset is sampled with probability proportional to the area of the itemset.…”

Section: Methodsmentioning

confidence: 99%

Mining Top-K Largest Tiles in a Data Stream

Lam

Pei

Prado

et al. 2014

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

Abstract. Large tiles in a database are itemsets with the largest area which is defined as the itemset frequency in the database multiplied by its size. Mining these large tiles is an important pattern mining problem since tiles with a large area describe a large part of the database. In this paper, we introduce the problem of mining top-k largest tiles in a data stream under the sliding window model. We propose a candidatebased approach which summarizes the data stream and produces the top-k largest tiles efficiently for moderate window size. We also propose an approximation algorithm with theoretical bounds on the error rate to cope with large size windows. In the experiments with two real-life datasets, the approximation algorithm is up to hundred times faster than the candidate-based solution and the baseline algorithms based on the state-of-the-art solutions. We also investigate an application of large tile mining in computer vision and in emerging search topics monitoring.

show abstract

Section: Methodsmentioning

confidence: 99%

Mining Top-K Largest Tiles in a Data Stream

Lam

Pei

Prado

et al. 2014

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

show abstract

“…In general, we can use the result of standard frequent pattern mining [2,20] although this incurs a high computational cost. Instead, we can resort to pattern sampling techniques [12,4], yet then we have to choose the number of patterns to be sampled. Alternatively, we [22] proposed to mine such pattern sets by the Minimum Description Length principle [11].…”

Section: Related Workmentioning

confidence: 99%

Efficiently Discovering Unexpected Pattern-Co-Occurrences

Bertens¹,

Vreeken

Siebes³

2017

Proceedings of the 2017 SIAM International Conference on Data Mining

View full text Add to dashboard Cite

Our world is filled with both beautiful and brainy people, but how often does a Nobel Prize winner also wins a beauty pageant? Let us assume that someone who is both very beautiful and very smart is more rare than what we would expect from the combination of the number of beautiful and brainy people. Of course there will still always be some individuals that defy this stereotype; these beautiful brainy people are exactly the class of anomaly we focus on in this paper. They do not posses intrinsically rare qualities, it is the unexpected combination of factors that makes them stand out.In this paper we define the above described class of anomaly and propose a method to quickly identify them in transaction data. Further, as we take a pattern set based approach, our method readily explains why a transaction is anomalous. The effectiveness of our method is thoroughly verified with a wide range of experiments on both real world and synthetic data.

show abstract

“…Boley et al [2,3] proposed to use Metropolis-Hastings sampling for the construction of data mining systems that do not require any userspecified threshold, i.e., minsup or minconf . However, all the algorithms generate approximate results and the completeness cannot be guaranteed.…”

Section: Related Workmentioning

confidence: 99%

ALPINE: Progressive Itemset Mining with Definite Guarantees

Imieliński²

2017

Proceedings of the 2017 SIAM International Conference on Data Mining

View full text Add to dashboard Cite

With increasing demand for efficient data analysis, execution time of itemset mining becomes critical for many large-scale or time-sensitive applications. We propose a dynamic approach for itemset mining that allows us to achieve flexible trade-offs between efficiency and completeness. ALPINE is to our knowledge the first algorithm to progressively mine itemsets and closed itemsets "support-wise". It guarantees that all itemsets with support exceeding the current checkpoint's support have been found before it proceeds further. Thus, it is very attractive for extremely long mining tasks with very high dimensional data because it can offer intermediate meaningful and complete results. This feature is the most important contribution of ALPINE, which is also fast but not necessarily the fastest algorithm around. Another critical advantage of ALPINE is that it does not require the apriori decided minimum support threshold.

show abstract

Direct local pattern sampling by efficient two-step random procedures

Cited by 59 publications

References 31 publications

Mining Top-K Largest Tiles in a Data Stream

Mining Top-K Largest Tiles in a Data Stream

Efficiently Discovering Unexpected Pattern-Co-Occurrences

ALPINE: Progressive Itemset Mining with Definite Guarantees

Contact Info

Product

Resources

About