An architecture and algorithms for multi-run clustering

Abstract

Abstract-This paper addresses two main challenges for clustering which require extensive human effort: selecting appropriate parameters for an arbitrary clustering algorithm and identifying alternative clusters. We propose an architecture and a concrete system MR-CLEVER for multi-run clustering that integrates active learning with clustering algorithms. The key hypothesis of this work is that better clustering results can be obtained by combining clusters that originate from multiple runs of clustering algorithms. By defining states that represent parameter settings of a clustering algorithm, the proposed architecture actively learns a state utility function. The utility of a parameter setting is assessed based on clustering run-time, quality and novelty of the obtained clusters. Furthermore, the utility function plays an important role in guiding the clustering algorithm to seek novel solutions. Cluster novelty measures are introduced for this purpose. Finally, we also contribute a cluster summarization algorithm that assembles a final clustering as a combination of high-quality clusters originating from multiple runs. Merits of our proposed system are that it is generic and therefore can be used in conjunction with different clustering algorithms, and it reduces human effort for selecting the parameters, for comparing clustering results and for assembling clustering results. We evaluate the proposed system in conjunction with a representative based clustering algorithm namely CLEVER for a challenging data mining task involving an earthquake dataset. The obtained results demonstrate that, in comparison to the best single-run clustering, multi-run clustering discovers solutions of higher quality.

A Framework for Multi-Objective Clustering and Its Application to Co-Location Mining

“…For this experiment we used a water well dataset called Arsenic_10_avg [9] which was created from a database provided by the Texas Water Development Board (TWDB) [10]. In this paper, we use a subset of this dataset containing 3 spatial attributes longitude, latitude and aquifer and the following 8 chemical concentrations for each water well: Arsenic (As), Molybdenum (Mo), Vanadium (V), Boron (B), Fluoride (F -), Chloride (Cl -), Sulfate (SO 4 2-) and Total Dissolved Solids (TDS). We used CLEVER (CLustEring using representatives and Randomized hill climbing), introduced in [5], as the clustering algorithm in the experiments.…”

“…Our clustering approach to cope with multi-objective problem is similar to the ensemble clustering approach as it is based on the key hypothesis that better clustering results can be obtained by combining clusters that originate from multiple runs of clustering algorithms [4]. However, our approach is an incremental approach that collects and refines clusters on the fly, and the search for alternative clusters takes into consideration what clusters already have been generated, rewarding novelty.…”

Retaining Customers by Data Mining: A Telecomunication Carrier's Case Study in China

“…The result of voice services using segmentation and data services using segmentation among customers in Group 1are presented by figure1 and figure 2 [3] [5]. We conduct a further segmentation among customers in Group 1 as follows mainly in accordance with the consuming characteristics, which covers 98% of the total population while the rest has no obvious characteristics.…”

A centralized adaptive hybrid routing mechanism for Ad Hoc networks

“…To further demonstrate the advantages of our mechanism, several simulation scenarios are designed in OPNET sim ulation platform [13]. As hybird routing protocols are most suitable for the large scale network with high node density, the principle of the scenario design is to compare the whole With simulations of different number of nodes, we have proved that in random moving and big scale scenarios, C AHR performs better in end-to-end latency, routing overhead, throughput and packet delivery ratio [14] [15].…”