Efficient clustering algorithms, such as k -Means, are often used in practice because they scale well for large datasets. However, they are only able to detect simple data characteristics. Ensemble clustering can overcome this limitation by combining multiple results of efficient algorithms. However, analysts face several challenges when applying ensemble clustering, i. e., analysts struggle to (a) efficiently generate an ensemble and (b) combine the ensemble using a suitable consensus function with a corresponding hyperparameter setting. In this paper, we propose EffEns, an efficient ensemble clustering approach to address these challenges. Our approach relies on meta-learning to learn about dataset characteristics and the correlation between generated base clusterings and the performance of consensus functions. We apply the learned knowledge to generate appropriate ensembles and select a suitable consensus function to combine their results. Further, we use a state-of-the-art optimization technique to tune the hyperparameters of the selected consensus function. Our comprehensive evaluation on synthetic and real-world datasets demonstrates that EffEns significantly outperforms state-of-the-art approaches w.r.t. accuracy and runtime.

Analysts often struggle with the combined algorithm selection and hyperparameter optimization problem, a.k.a. CASH problem in literature. Typically, they execute several algorithms with varying hyperparameter settings to find configurations that show valuable results. Efficiently finding these configurations is a major challenge. In clustering analyses, analysts face the additional challenge to select a cluster validity index that allows them to evaluate clustering results in a purely unsupervised fashion. Many different cluster validity indices exist and each one has its benefits depending on the dataset characteristics. While experienced analysts might address these challenges using their domain knowledge and experience, especially novice analysts struggle with them. In this paper, we propose a new meta-learning approach to address these challenges. Our approach uses knowledge from past clustering evaluations to apply strategies that experienced analysts would exploit. In particular, we use meta-learning to (a) select a suitable clustering validity index, (b) efficiently select well-performing clustering algorithm and hyperparameter configurations, and (c) reduce the search space to suitable clustering algorithms. In the evaluation, we show that our approach significantly outperforms state-of-the-art approaches regarding accuracy and runtime.

Abstract Clustering is a fundamental primitive in manifold applications. In order to achieve valuable results in exploratory clustering analyses, parameters of the clustering algorithm have to be set appropriately, which is a tremendous pitfall. We observe multiple challenges for large-scale exploration processes. On the one hand, they require specific methods to efficiently explore large parameter search spaces. On the other hand, they often exhibit large runtimes, in particular when large datasets are analyzed using clustering algorithms with super-polynomial runtimes, which repeatedly need to be executed within exploratory clustering analyses. We address these challenges as follows: First, we present LOG-Means and show that it provides estimates for the number of clusters in sublinear time regarding the defined search space, i.e., provably requiring less executions of a clustering algorithm than existing methods. Second, we demonstrate how to exploit fundamental characteristics of exploratory clustering analyses in order to significantly accelerate the (repetitive) execution of clustering algorithms on large datasets. Third, we show how these challenges can be tackled at the same time. To the best of our knowledge, this is the first work which simultaneously addresses the above-mentioned challenges. In our comprehensive evaluation, we unveil that our proposed methods significantly outperform state-of-the-art methods, thus especially supporting novice analysts for exploratory clustering analyses in large-scale exploration processes.

\n Data analysis is a highly iterative process. In order to achieve\n valuable analysis results, analysts typically execute many\n configurations, i.e., algorithms and their hyperparameter settings,\n based on their domain knowledge. While experienced analysts may be\n able to define small search spaces for promising configurations,\n especially novice analysts define large search spaces due to their\n lack of domain knowledge. In the worst case, they perform an\n exhaustive search throughout the whole search space, resulting in\n infeasible runtimes. Recent advances in the research area of AutoML\n address this challenge by supporting novice analysts in the combined\n algorithm selection and hyperparameter optimization (CASH) problem\n for supervised learning tasks. However, no such systems exist for\n unsupervised learning tasks, such as the prevalent task of\n clustering analysis. In this work, we present our novel AutoML4Clust\n approach, which efficiently supports novice analysts regarding CASH\n for clustering analyses. To the best of our knowledge, this is the\n first thoroughly elaborated approach in this area. Our comprehensive\n evaluation unveils that AutoML4Clust significantly outperforms\n several existing approaches, as it achieves considerable speedups\n for the CASH problem, while still achieving very valuable clustering\n results.\n

The plethora of available analytic platforms escalates the difficulty of selecting the most appropriate platform for a certain data mining task and datasets with varying characteristics. Especially novice analysts experience difficulties to keep up with the latest technical developments. In this demo, we present the ASAP-DM framework. ASAP-DM is able to automatically select a well-performing analytic platform for a given data mining task via an intuitive web interface, thus especially supporting novice analysts. The take-aways for demo attendees are: (1) a good understanding of the challenges of various data mining workloads, dataset characteristics, and the effects on the selection of analytic platforms, (2) useful insights on how ASAP-DM internally works, and (3) how to benefit from ASAP-DM for exploratory data analysis.