A Surrogate-Assisted Multiobjective Evolutionary Algorithm for Large-Scale Task-Oriented Pattern Mining

Despite the success of Evolutionary Algorithms in solving complex problems, they may require many function evaluations. This becomes an issue when dealing with costly problems. Surrogate models may overcome this difficulty, though their use in problems with medium to large dimensionality is underexplored in the literature. Problems with multiple conflicting objectives can be formulated as Multi-objective Optimization Problems (MOPs). MOPs have received a great attention lately, mainly with Multi-objective Evolutionary Algorithms (MOEAs). This paper proposes ELMOEA/D-DE, a surrogate-assisted MOEA, for solving expensive MOPs in small evaluation budgets. ELMOEA/D-DE encompasses a state-of-the-art MOEA based on decomposition, Differential Evolution (DE) operators and Extreme Learning Machines. This paper tests three variants of ELMOEA/D-DE, using different DE operators, for solving five known benchmark MOPs with 10 to 60 decision variables. All variants achieve good results in terms of hyper volume metric and the best variant with operator DE/rand/1/bin is compared with two state-of-the-art approaches (MOEA/D-RBF and a non-surrogate-based MOEA), achieving the best results in all but one problems instances.

High-dimensional expensive optimization by Kriging-assisted multiobjective evolutionary algorithm with dimensionality reduction

Surrogate-assisted multi-objective evolutionary feature selection of generation-based fixed evolution control for time series forecasting with LSTM networks

Closed frequent similar pattern mining: Reducing the number of frequent similar patterns without information loss

Adaptive surrogate-assisted multi-objective evolutionary algorithm using an efficient infill technique

Enhancing the maintenance strategy and cost in systems with surrogate assisted multiobjective evolutionary algorithms

Frequent pattern mining is a classic data mining technique, generally applicable to a wide range of application domains, and a mature area of research. The fundamental challenge arises from the combinatorial nature of frequent itemsets, scaling exponentially with respect to the number of unique items. Apriori-based and FPTree-based algorithms have dominated the space thus far. Initial phases of this research relied on the Apriori algorithm and utilized a distributed computing environment; we proposed the Cartesian Scheduler to manage Apriori's candidate generation process. To address the limitation of bottom-up frequent pattern mining algorithms such as Apriori and FPGrowth, we propose the Frequent Hierarchical Pattern Tree (FHPTree): a tree structure and new frequent pattern mining paradigm. The classic problem is redefined as frequent hierarchical pattern mining where the goal is to detect frequent maximal pattern covers. Under the proposed paradigm, compressed representations of maximal patterns are mined using a top-down FHPTree traversal, FHPGrowth, which detects large patterns before their subsets, thus yielding significant reductions in computation time. The FHPTree memory footprint is small; the number of nodes in the structure scales linearly with respect to the number of unique items. Additionally, the FHPTree serves as a persistent, dynamic data structure to index frequent patterns and enable efficient searches. When the search space is exponential, efficient targeted mining capabilities are paramount; this is one of the key contributions of the FHPTree. This dissertation will demonstrate the performance of FHPGrowth, achieving a 300x speed up over state-of-the-art maximal pattern mining algorithms and approximately a 2400x speedup when utilizing FHPGrowth in a distributed computing environment. In addition, we allude to future research opportunities, and suggest various modifications to further optimize the FHPTree and FHPGrowth. Moreover, the methods we offer will have an impact on other data mining research areas including contrast set mining as well as spatial and temporal mining.

Recommender systems are chiefly renowned for their applicability in e-commerce sites and social media. For system optimization, this work introduces a method of behaviour pattern mining to analyze the person's mental stability. With the utilization of the sequential pattern mining algorithm, efficient extraction of frequent patterns from the database is achieved. A candidate sub-sequence generation-and-test method is adopted in conventional sequential mining algorithms like the Generalized Sequential Pattern Algorithm (GSP). However, since this approach will yield a huge candidate set, it is not ideal when a large amount of data is involved from the social media analysis. Since the data is composed of numerous features, all of which may not have any relation with one another, the utilization of feature selection helps remove unrelated features from the data with minimal information loss. In this work, Frequent Pattern (FP) mining operations will employ the Systolic tree. The systolic tree-based reconfigurable architecture will offer various benefits such as high throughput as well as cost-effective performance. The database's frequently occurring item sets can be found by using the FP mining algorithms. Numerous research areas related to machine learning and data mining are fascinated by feature selection since it will enable the classifiers to be swift, more accurate, and cost-effective. Over the last ten years or so, there have been significant technological advancements in heuristic techniques. These techniques are beneficial because they improve the search procedure's efficiency, albeit at the potential sacrifice of completeness claims. A new recommender system for mental illness detection was based on features selected using River Formation Dynamics (RFD), Particle Swarm Optimization (PSO), and hybrid RFD-PSO algorithm is proposed in this paper. The experiments use the depressive patient datasets for evaluation, and the results demonstrate the improved performance of the proposed technique.

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>

The Substantial amount of research has been done in the area of frequent pattern mining in the last few decades. Researchers have developed various algorithms to generate frequent patterns. We propose Pattern Mining using Linked list (PML) algorithm that generates frequent patterns using Linked list. It uses both horizontal and vertical data layout. To generate 1-itemsets, it uses horizontal data layout and for 2-itemsets and more, it uses vertical data layout. The important feature of vertical data layout is that it count the frequency fast using intersection operations on transaction ids (tids). It prunes automatically irrelevant data. The algorithm uses Linked list data structure due to which it takes less execution time to generate frequent patterns. It runs with efficient memory usage. It scans the dataset only two times. The experimental results of proposed algorithm have been compared with other algorithms.

Frequent pattern mining is a basic problem in data mining and knowledge discovery. The discovered patterns can be used as the input for analyzing association rules, mining sequential patterns, recognizing clusters, and so on. However, discovering frequent patterns in large scale datasets is an extremely time consuming task. Most research in the area of association rule discovery has focused on the method of efficient frequent pattern discovery, e.g. Park, Chen & Yu (1995); Savasere, Omiecinski & Navathe (1995); Han, Pei & Yin (2000); Pei, Han, Lu, Nishio, Tang & Yang (2001). When seeking all associations that satisfy constraints on support and confidence, once frequent patterns have been identified, generating the association rules is trivial. In the last decade, various algorithms have been proposed on this problem, e.g. maximal pattern mining - Grahne & Zhu (2003); closed pattern mining - Pei, Han & Mao (2000); Grahne & Zhu (2003); mining the most interesting frequent patterns - Fu, Kwong & Tang (2000); Han, Wang, Lu & Tzvetkov (2002, 2005); Hirate, Iwahashi & Yamana (2004). However, some challenges are still existed and need to be overcome. In this paper, a mathematical space will be introduced with some new related concepts and propositions to design a new algorithm for solving frequent patterns mining problem. It is hoped that such an improved algorithm will be simple to implement and more efficient.

Real life sequential databases are usually not static. They tend to grow incrementally. So after every update a frequent pattern may no longer remains frequent while some infrequent patterns may appear as frequent in updated database. It is not a good idea to mine sequential database from scratch every time as the update occurs. It would be better if we can use the knowledge of already mined sequential patterns to find the complete set of sequential patterns for updated database. An incremental mining algorithm does the same thing. The main goal of an incremental mining algorithm is to reduce the time taken to find out the frequent patterns significantly i.e. it should mine the set of frequent patterns in significantly less time than a non-incremental mining algorithm. In this work we have approached using weight constraints, in time and space, of an idea of already existing algorithm called WSM.

High volumes of wide varieties of valuable data of different veracities can be easily generated or collected at a high velocity from various big data applications and services. Embedded in these big data are valuable knowledge and useful information, which can be discovered by data science solutions. As a popular data science task, frequent pattern mining aims to discover implicit, previously unknown and potentially useful information and valuable knowledge in terms of sets of frequently co-occurring items. Many of the existing frequent pattern mining algorithms return large numbers of frequent patterns, of which only a small portion may be of interest to users. In this paper, we present a constrained mining algorithm that allows crowds of users to collaboratively vote for their interesting patterns. Such an algorithm takes the benefits of crowdsourcing, crowdvoting and collaborative filtering for the data analytics and mining of popular constrained frequent patterns from big data applications and services.

Pattern mining refers to a subfield of data mining that uncovers interesting, unexpected, and useful patterns from transaction databases. Such patterns reflect frequent and infrequent patterns. An abundant literature has dedicated in frequent pattern mining and tremendous efficient algorithms for frequent itemset mining in the transaction database. Nonetheless, the infrequent pattern mining has emerged to be an interesting issue in discovering patterns that rarely occur in the transaction database. More researchers reckon that rare pattern occurrences may offer valuable information in knowledge data discovery process. The R-Eclat is a novel algorithm that determines infrequent patterns in the transaction database. The multiple variants in the R-Eclat algorithm generate varied performances in infrequent mining patterns. This paper proposes IF-Postdiffset as a new variant in R-Eclat algorithm. This paper also highlights the performance of infrequent mining pattern from the transaction database among different variants of the R-Eclat algorithm regarding its execution time.

Most algorithms for frequent pattern mining use a support-based pruning strategy to prune a combinatorial search space. However, they are not effective for finding correlated patterns with similar levels of support. In additional, traditional patterns mining algorithms rarely consider weighted pattern mining. In this paper, we present a new algorithm, WHFPMiner (Weighted Highly-correlated Frequent Patterns Miner) in which a new objective measure, called weighted h-confidence, is developed to mine weighted highly-correlated frequent patterns with similar levels of weighted support. Adopting an improved weighted FP-tree structure, this algorithm exploits both cross-weighted support and anti-monotone properties of the weighted h-confidence measure for the efficient discovery of weighted hyperclique patterns. A comprehensive performance study shows that WHFPMiner is efficient and fast for finding weighted highly-correlated frequent patterns. Moreover, it generates fewer but more valuable patterns with the high correlation.KeywordsData miningWeighted h-confidenceHighly-correlatedFrequent patternFP-tree