Effects of Population Initialization on Evolutionary Techniques for Subgroup Discovery in High Dimensional Datasets

MEFASD-BD: Multi-objective evolutionary fuzzy algorithm for subgroup discovery in big data environments - A MapReduce solution

Evolutionary algorithms for subgroup discovery in e-learning: A practical application using Moodle data

This paper presents an evolutionary approach for mining diverse and more informative subgroups focused on high dimensional data sets. Subgroup Discovery (SD) is an important tool for knowledge discovery that aims to identify sets of features that distinguish a target group from the others (e.g. successful from unsuccessful treatments). At the same time, to extract information from high dimensional data sets becomes more natural. One of the first and most efficient SD heuristics focused on high dimensional data is the SSDP. However, this model deals superficially with diverse/redundancy in top-k subgroups, which can result in poor information for users. This work presents SSDP+, an extension of the SSDP model to provide diversity in a way that explore the relation between subgroups order to 2enerate a more informative set of patterns.

Metaheuristic algorithms have been hybridized with the standard K-means to address the latter’s challenges in finding a solution to automatic clustering problems. However, the distance calculations required in the standard K-means phase of the hybrid clustering algorithms increase as the number of clusters increases, and the associated computational cost rises in proportion to the dataset dimensionality. The use of the standard K-means algorithm in the metaheuristic-based K-means hybrid algorithm for the automatic clustering of high-dimensional real-world datasets poses a great challenge to the clustering performance of the resultant hybrid algorithms in terms of computational cost. Reducing the computation time required in the K-means phase of the hybrid algorithm for the automatic clustering of high-dimensional datasets will inevitably reduce the algorithm’s complexity. In this paper, a preprocessing phase is introduced into the K-means phase of an improved firefly-based K-means hybrid algorithm using the concept of the central limit theorem to partition the high-dimensional dataset into subgroups of randomly formed subsets on which the K-means algorithm is applied to obtain representative cluster centers for the final clustering procedure. The enhanced firefly algorithm (FA) is hybridized with the CLT-based K-means algorithm to automatically determine the optimum number of cluster centroids and generate corresponding optimum initial cluster centroids for the K-means algorithm to achieve optimal global convergence. Twenty high-dimensional datasets from the UCI machine learning repository are used to investigate the performance of the proposed algorithm. The empirical results indicate that the hybrid FA-K-means clustering method demonstrates statistically significant superiority in the employed performance measures and reducing computation time cost for clustering high-dimensional dataset problems, compared to other advanced hybrid search variants.

Pattern mining is considered as a really interesting task for the extraction of hidden knowledge in the form of patterns. The extraction of such subsequences, substructures or itemsets that represent any type of homogeneity and regularity in data has been carried out from unlabeled data. However, there are many research areas that aim at discovering patterns in the form of rules induced from labeled data. Hence, it is interesting to discover patterns and associations from a supervised point of view since a single item or a set of them can be considered as distinctive. This task, which is known as supervised local pattern mining, is described in this chapter, including different areas such as contrast set mining, emerging pattern mining, and subgroup discovery. This chapter is mainly focused on the subgroup discovery task, which is widely known in the field of supervised local pattern mining. Here, an exhaustive description about this task is provided, including some important quality measures in the field. Then, this chapter includes different evolutionary approaches for subgroup discovery. Finally, this chapter includes an analysis of other different approaches proposed for mining and quantifying local patterns.

This paper presents a genetic fuzzy system for the data mining task of subgroup discovery, the subgroup discovery iterative genetic algorithm (SDIGA), which obtains fuzzy rules for subgroup discovery in disjunctive normal form. This kind of fuzzy rule allows us to represent knowledge about patterns of interest in an explanatory and understandable form that can be used by the expert. Experimental evaluation of the algorithm and a comparison with other subgroup discovery algorithms show the validity of the proposal. SDIGA is applied to a market problem studied in the University of Mondragon, Spain, in which it is necessary to extract automatically relevant and interesting information that helps to improve fair planning policies. The application of SDIGA to this problem allows us to obtain novel and valuable knowledge for experts.

Subgroup discovery is a learning task that aims at finding interesting rules from classified examples. The search is guided by a utility function, trading off the coverage of rules against their statistical unusualness. One shortcoming of existing approaches is that they do not incorporate prior knowledge. To this end a novel generic sampling strategy is proposed. It allows to turn pattern mining into an iterative process. In each iteration the focus of subgroup discovery lies on those patterns that are unexpected with respect to prior knowledge and previously discovered patterns. The result of this technique is a small diverse set of understandable rules that characterise a specified property of interest. As another contribution this article derives a simple connection between subgroup discovery and classifier induction. For a popular utility function this connection allows to apply any standard rule induction algorithm to the task of subgroup discovery after a step of stratified resampling. The proposed techniques are empirically compared to state of the art subgroup discovery algorithms.

Supervised descriptive rule discovery techniques like subgroup discovery are quite popular in applications like fraud detection or clinical studies. Compared with other descriptive techniques, like classical support/confidence association rules, subgroup discovery has the advantage that it comes up with only the top-k patterns, and that it makes use of a quality function that avoids patterns uncorrelated with the target. If these techniques are to be applied in privacy-sensitive scenarios involving distributed data, precise guarantees are needed regarding the amount of information leaked during the execution of the data mining. Unfortunately, the adaptation of secure multi-party protocols for classical support/confidence association rule mining to the task of subgroup discovery is impossible for fundamental reasons. The source is the different quality function and the restriction to a fixed number of patterns – i.e. exactly the desired features of subgroup discovery. In this paper, we present a new protocol which allows distributed subgroup discovery while avoiding the disclosure of the individual databases. We analyze the properties of the protocol, describe a prototypical implementation and present experiments that demonstrate the feasibility of the approach.KeywordsAssociation RuleQuality FunctionPrivacy PreserveSubgroup DiscoveryPrivacy Preserve Data MiningThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Surrogate-assisted evolutionary algorithms (SAEAs) have been widely developed to solve complex and computationally expensive multiobjective optimization problems (EMOPs) in recent years. However, when dealing with high-dimensional optimization problems in decision space, the performance of these surrogate-assisted multiobjective evolutionary algorithms (MOEAs) deteriorates drastically. In this work, a novel classifier-assisted rank-based learning and local model-based multiobjective evolutionary algorithm (CLMEA) is proposed for high-dimensional EMOPs. CLMEA makes full use of the uncertainty of solutions in the decision space and objective space to explore the uncertain but informative space toward high-dimensional problems. Specifically, the offspring in different ranks uses rank-based learning strategy to generate more promising and informative candidates for function evaluations (FEs). To reduce the search region of high-dimensional problems and maintain the diversity of solutions, the most uncertain sample point from the nondominated solutions measured by the crowding distance is selected as the center to conduct local search. The experimental results of benchmark problems and a real-world application on geothermal reservoir heat extraction optimization demonstrate superior performance of CLMEA compared with the state-of-the-art surrogate-assisted MOEAs. The source code for this work is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/JellyChen7/CLMEA</uri>

For the task of subgroup discovery, generalization-aware interesting measures that are based not only on the statistics of the patterns itself, but also on the statistics of their generalizations have recently been shown to be essential. A key technique to increase runtime performance of subgroup discovery algorithms is the application of optimistic estimates to limit the search space size. These are upper bounds for the interestingness that any specialization of the currently evaluated pattern may have. Until now these estimates are based on the anti-monotonicity of instances, which are covered by the current pattern. This neglects important properties of generalizations. Therefore, we present in this paper a new scheme of deriving optimistic estimates for generalization aware subgroup discovery, which is based on the instances by which patterns differ in comparison to their generalizations. We show, how this technique can be applied for the most popular interestingness measures for binary as well as for numeric target concepts. The novel bounds are incorporated in an efficient algorithm, which outperforms previous methods by up to an order of magnitude.KeywordsAssociation RuleQuality FunctionOptimistic EstimateTarget ConceptPositive InstanceThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Evolutionary algorithms for subgroup discovery usually randomly initialize the population, which often causes them to spend part of their time evaluating unpromising solutions. This situation causes the algorithm to take more time to converge to good solutions. In this paper, we present a new initial population construction heuristic for DINOS, a genetic subgroup discovery algorithm that mines non-redundant subgroups with high quality in a short time. The proposed heuristic is based on the generation of a collection of decision trees, allowing to obtain an initial population in which all the rules are valid and with a large coverage of the database. The quality of these rules is also high and they contain a large diversity in the attributes used, allowing to deal with problems having a large number of dimensions. The experiments carried out show that the new method allows mining more high-quality and diverse subgroups in a slightly higher computational time.

To overcome curse of dimensionality for outlier detecting in high dimensional dataset, axis-parallel subspace (SOD) and angle-based outlier detection (ABOD) methods were presented. These methods are also friendly used distance-based to detect outliers. In this regard, based on the reality of fuzzy data for explaining the world phenomena, this paper introduces an extended version of both methods for fuzzy dataset. First, the basic concepts of both methods are explained. Next we provide two metrics based on Euclidean and analytic distance to measure distance between fuzzy objects; also Cosine similarity measure formula for calculating the cosine of angle between two difference vectors in high-dimensional fuzzy dataset is illustrated. Then the algorithms to determine outliers of fuzzy datasets by using these metrics and Cosine similarity measure, based on ABOD and SOD algorithms, are presented. Some numerical experimental examples are also presented, in which both real and synthesis datasets are used, For a real numerical examination, we have applied proposed algorithms to data from 15 Iranian petrochemical companies in a fully fuzzy environment. The obtained results show the significant properties of the new methods in detecting outliers.

In data mining, the process of data obtained from users history databases is called Web usage mining. The main benefits lie in the improvement of the design of Web applications for the final user. This paper presents the application of subgroup discovery (SD) algorithms based on evolutionary fuzzy systems (EFSs) to the data obtained in an e-commerce Web site of extra virgin olive oil sale called http://www.orolivesur.com. For this purpose, a brief description of the SD process (objectives, properties, quality measures) and EFSs is presented. A discussion about the results obtained will also be included, especially focusing on the interests of the designer team of the Web site, providing some guidelines for improving several aspects such as usability and user satisfaction.

This paper presents a multiobjective genetic algorithm which obtains fuzzy rules for subgroup discovery in disjunctive normal form. This kind of fuzzy rules lets us represent knowledge about patterns of interest in an explanatory and understandable form which can be used by the expert. The evolutionary algorithm follows a multiobjective approach in order to optimize in a suitable way the different quality measures used in this kind of problems. Experimental evaluation of the algorithm, applying it to a market problem studied in the University of Mondragón (Spain), shows the validity of the proposal. The application of the proposal to this problem allows us to obtain novel and valuable knowledge for the experts.

Identifying complex interactions among millions of single nucleotide polymorphisms (SNPs) is a key challenge in Genome-Wide Association Studies (GWAS), offering crucial insights into the genetic architecture of complex diseases. Evolutionary algorithm (EA)-based methods have gained significant attention for their global search capabilities, controllable runtime, and multi-objective optimization potential. However, when applied to high-dimensional GWAS datasets, many existing EA-based methods encounter challenges such as getting trapped in local optima and facing high computational demands. To address these issues, the evolutionary multitasking (EMT) paradigm presents a promising solution, enhancing population diversity and convergence speed through collaborative, cross-task knowledge sharing. Furthermore, the multi-tasking framework and EA can be seamlessly deployed across multiple Graphics Processing Units (GPUs), leveraging their high parallelism and aggregated memory bandwidth. Therefore, we introduce a GPU-powered evolutionary auxiliary multitasking algorithm (GEAMT) for fast SNP interaction detection. GEAMT first constructs a main task along with several low-dimensional auxiliary tasks to redefine the original task. The main task explores the entire search space, while the auxiliary tasks search distinct subspaces to enhance local optimization capabilities. In each iteration, the auxiliary tasks transfer high-quality information to the main task via an information transfer mechanism. Subsequently, an auxiliary task update strategy based on feature regrouping is employed to switch the search subspaces of the auxiliary tasks. The final results are derived from the Pareto-optimal solutions of the main task. Implemented across multiple GPUs, GEAMT achieves notable scalability and efficiency. Comprehensive experiments on both synthetic and real-world datasets demonstrate that GEAMT can significantly enhance search accuracy and speed up the search process.