Frequent Pattern Mining Algorithm Research Articles

Measuring the Significance of Spatiotemporal Co-Occurrences

Spatiotemporal co-occurrences are the appearances of spatial and temporal overlap relationships among trajectory-based spatiotemporal instances with region-based geometric representations. Assessing the significance of spatiotemporal co-occurrences plays an important role in the spatiotemporal frequent pattern mining applications of moving region objects. A spatiotemporal version of the popular Jaccard measure has been used for measuring the strength of spatiotemporal co-occurrences. We will demonstrate the shortcomings of the Jaccard ( J ) measure when it is used for assessing the significance of co-occurrences among spatiotemporal instances with highly different spatiotemporal evolution characteristics. We will present two extended novel measures ( J + and J * ) that address the problems linked to the J measure. Our work includes algorithms for the significance measure calculations, the proofs and explanations about the key properties of measures, and a detailed experimental evaluation section. Our experiments include in-depth relevancy and running time analyses demonstrating the suitability of our proposed measures for spatiotemporal frequent pattern mining algorithms.

A New Frequent Pattern Mining Algorithm with Weighted Multiple Minimum Supports

Association rules mining is one of the momentous areas in data mining. Frequent patterns mining plays an important role in association rules mining. The effects of traditional frequent patterns min...

MapReduce Based Multilevel Consistent and Inconsistent Association Rule Detection from Big Data Using Interestingness Measures

Abstract Multilevel association rule mining in distributed environment plays an important role in big data analysis for making marketing strategy. Multilevel association rule provides more significant information than single level rule, and also discovers the conceptual hierarchy of knowledge from the hierarchical dataset. In this era of internet, various online marketing sites and social networking sites are generating enormous amount of data so that it becomes very difficult to process and analyze it using conventional approaches as it consumes more time. This paper overcomes the computing limitation of single node by distributing the task on multi-node cluster. The proposed method initially extracts multilevel association rules including level-crossing for each zone using distributed multilevel frequent pattern mining algorithm (DMFPM). These generated multilevel association rules are so large that it becomes complex to analyze it. Thus, MapReduce based multilevel consistent and inconsistent rule detection (MR-MCIRD) algorithm is proposed to detect the consistent and inconsistent multilevel rules from big hierarchical data which provide useful and actionable knowledge to the domain experts. These pruned interesting rules also give useful knowledge for better marketing strategy. The extracted multilevel consistent and inconsistent rules are evaluated and compared based on different interestingness measures presented together with experimental results that lead to the final conclusions.

Interesting association rule mining with consistent and inconsistent rule detection from big sales data in distributed environment

Nowadays, there is an increasing demand in mining interesting patterns from the big data. The process of analyzing such a huge amount of data is really computationally complex task when using traditional methods. The overall purpose of this paper is in twofold. First, this paper presents a novel approach to identify consistent and inconsistent association rules from sales data located in distributed environment. Secondly, the paper also overcomes the main memory bottleneck and computing time overhead of single computing system by applying computations to multi node cluster. The proposed method initially extracts frequent itemsets for each zone using existing distributed frequent pattern mining algorithms. The paper also compares the time efficiency of Mapreduce based frequent pattern mining algorithm with Count Distribution Algorithm (CDA) and Fast Distributed Mining (FDM) algorithms. The association generated from frequent itemsets are too large that it becomes complex to analyze it. Thus, Mapreduce based consistent and inconsistent rule detection (MR-CIRD) algorithm is proposed to detect the consistent and inconsistent rules from big data and provide useful and actionable knowledge to the domain experts. These pruned interesting rules also give useful knowledge for better marketing strategy as well. The extracted consistent and inconsistent rules are evaluated and compared based on different interestingness measures presented together with experimental results that lead to the final conclusions.

DYNAMIC LOCATION AREA PLANNING IN CELLULAR NETWORK USING FREQUENT PATTERN MINING

Frequent pattern mining algorithms as the name says, mines sets frequent patterns form given datasets. These algorithms provide immensely helpful results which have a wide scope of application starting from simple decision problems to complex business intelligence aspects. This paper attempts to apply the same concept of frequent pattern mining to solve the location management problem of GSM networks. In GSM networks the task of keeping track of a mobile user (MU) and relaying an incoming call, is called location management. It basically includes two processes, location update and paging. Location update deals with managing the current location of the MUs. There are many approaches to do this like time based, movement based, distance based etc. In this paper the location update procedure relies on the collected data of user movements form one network cell to another cell. This data has a definite pattern, as in day to day life a person mostly has a fixed route of travelling e.g. home to office in the morning and office to home in the evening. During this movement he crosses a specific set of network cells which remains same throughout the week. Thus, frequent pattern mining algorithm can be applied on the user’s mobility log and try to find out the most probable location where the mobile user could be found. Using the results, a dynamic location area for individual user’s current location can be created. Thus minimizing cost related to location update which otherwise involves communication between the mobile handset and the base station, and calculations related to keeping track of the location of mobile users.

A fast and parallel algorithm for frequent pattern mining from big data in many-task environments

Many studies have tried to efficiently discover frequent patterns in large databases. The algorithms used in these studies fall into two main categories: apriori algorithms and frequent pattern growth (FP-growth) algorithms. Apriori algorithms operate according to a generate-and-test approach, so performance suffers from the testing of too many candidate itemsets. Therefore, most recent studies have applied an FP-growth approach to the discovery of frequent patterns. The rapid growth of data, however, has introduced new challenges for the mining of frequent patterns, in terms of both execution efficiency and scalability. Big data often contains a large number of items, a large number of transactions and long average transaction length, which result in large FP-trees. In addition to its dependence on data characteristics, FP-tree size is also sensitive to the minimum support threshold. This is because the small support is probable to bring many branches for nodes, greatly enlarging the FP-tree and the number of reconstructed conditional pattern-based trees. In this paper, we propose a novel algorithm and architecture for efficiently mining frequent patterns from big data in distributed many-task computing environments. Through empirical evaluation of various simulation conditions, we show that the proposed method delivers excellent execution time.

A fast and parallel algorithm for frequent pattern mining from big data in many-task environments

Many studies have tried to efficiently discover frequent patterns in large databases. The algorithms used in these studies fall into two main categories: apriori algorithms and frequent pattern growth (FP-growth) algorithms. Apriori algorithms operate according to a generate-and-test approach, so performance suffers from the testing of too many candidate itemsets. Therefore, most recent studies have applied an FP-growth approach to the discovery of frequent patterns. The rapid growth of data, however, has introduced new challenges for the mining of frequent patterns, in terms of both execution efficiency and scalability. Big data often contains a large number of items, a large number of transactions and long average transaction length, which result in large FP-trees. In addition to its dependence on data characteristics, FP-tree size is also sensitive to the minimum support threshold. This is because the small support is probable to bring many branches for nodes, greatly enlarging the FP-tree and the number of reconstructed conditional pattern-based trees. In this paper, we propose a novel algorithm and architecture for efficiently mining frequent patterns from big data in distributed many-task computing environments. Through empirical evaluation of various simulation conditions, we show that the proposed method delivers excellent execution time.

Enabling Far-Edge Analytics: Performance Profiling of Frequent Pattern Mining Algorithms

Far-edge analytics refers to the enablement of data mining algorithms in far-edge mobile devices that are part of mobile edge cloud computing (MECC) systems. Far-edge analytics enables data reduction in mobile environments, hence reducing the data transfer rate and bandwidth utilization cost for mobile-edge communication. In addition, far-edge analytics facilitates local knowledge availability to enable personalized mobile data stream mining applications. Existing literature mainly addresses classification and clustering problems in far-edge mobile devices, but the problem of frequent pattern mining (FPM) remains unexplored. This paper presents the results of an experimental study on the performance profiling of frequent pattern mining algorithms. We developed a real mobile application for performance analysis and profiling of 21 FPM algorithms with various real data sets in terms of execution time, storage complexity, sparsity, density, and data set size. According to the experimental results, large-sized data sets with high sparsity increase computational and storage cost in far-edge mobile devices. To address these issues, we propose a framework and discuss the relevant research challenges for seamless execution of FPM algorithms in MECC systems.

Performance Comparison of Frequent Pattern Mining Algorithms for Business Intelligence Analytics

Objectives: In this paper, a simple and flexible partition algorithm has been proposed to mine frequent data item sets. This partition algorithm is different from other frequent pattern mining algorithm like Apriori algorithm, AprioriAllHybrid algorithm etc. Method: Partition algorithm concept has been proposed to increase the execution speed with minimum cost. Initially only for one time the database is scanned and separate partitions will be created for each sets of itemsets, which is 1-itemset, 2-itemsets, 3-itemsets etc. Findings: The scanning of whole database is not necessary to get the count of an itemset, it is enough to get the count of each data itemsets from its partition. This partition algorithm approach is implemented and evaluated against AprioriAllHybrid and Apriori algorithm. The candidate itemsets generated at each step is reduced and the scanning time is also reduced. The proposed methodology performance is significantly better than other algorithms and it promotes the faster execution time for mining frequent patterns. Applications: This proposed algorithm is used in areas like retail sales, production, universities, finance, banking systems and for business to plan and estimate the future values.

DPcode: Privacy-Preserving Frequent Visual Patterns Publication on Cloud

Nowadays, cloud has become a promising multimedia data processing and sharing platform. Many institutes and companies plan to outsource and share their large-scale video and image datasets on cloud for scientific research and public interest. Among various video applications, the discovery of frequent visual patterns over graphical data is an exploratory and important technique. However, the privacy concerns over the leakage of sensitive information contained in the videos/images impedes the further implementation. Although the frequent visual patterns mining (FVPM) algorithm aggregates summary over individual frames and seems not to pose privacy threat, the private information contained in individual frames still may be leaked from the statistical result. In this paper, we study the problem of privacy-preserving publishing of graphical data FVPM on cloud. We propose the first differentially private frequent visual patterns mining algorithm for graphical data, named DPcode. We propose a novel mechanism that integrates the privacy-preserving visual word conversion with the differentially private mechanism under the noise allocation strategy of the sparse vector technique. The optimized algorithms properly allocate the privacy budgets among different phases in FPM algorithm over images and reduce the corresponding data distortion. Extensive experiments are conducted based on datasets commonly used in visual mining algorithms. The results show that our approach achieves high utility while satisfying a practical privacy requirement.

Frequent Itemset Mining in Data Mining: A Survey

Data mining is process of extracting useful information from different perspectives. Frequent Item set mining is widely used in financial, retail and telecommunication industry. The major concern of these industries is faster processing of a very large amount of data. Frequent item sets are those items which are frequently occurred. So we can use different types of algorithms for this purpose. Frequent Item se performed Apriori, FP-tree, Eclat , and RARM the work in this paper, we have analyzed algorithms for finding frequent patterns with the purpose of discovering how these algorithms can be used to obtain frequent patterns over large transactional databases. This has been presented in the form of a compa rative study of the following algorithms: Apriori, Freque nt Pattern (FP) Growth Rapid Association Rule Mining (RARM) and ECLAT algorithm frequent pattern mining algorithms. This study also focuses on each of the algorithm's advantages, disadvantages and limitations for finding patterns among large item sets in database systems. Keywordsrule, Frequent Item, Data Mining

Novel Algorithms CIPFP for Mining Frequent Patterns using Counting Inference from Probabilistic Databases and Future Possibilities

We consider the problem of discovering frequent item sets and association rules between items in a large database of transactional databases acquired under uncertainty. A probabilistic database considered here is one in which with each transaction associated is a probability, represents the confidence that the transaction will occur with given associated certainty. In this paper, we address the problem of the efficiency of the main phase of most data mining applications: The frequent pattern extraction. This problem is mainly related to the number of operations required for counting pattern supports in the database and we propose a new method, called counting inference probabilistic frequent pattern miner in probabilistic databases, this algorithm allows to perform as few support counts as possible. It is optimized to reduce the number of database scan as well as the number of patterns for which explicit support count is required. Using this method, the support of a pattern is determined without accessing the database whenever possible, using the supports of some of its sub-patterns called key patterns. This method was implemented in the CIPFP, counting inference based probabilistic frequent pattern mining algorithm that is an optimization of the simple and efficient Apriori algorithm. The goal is to transform all key patterns into non-key patterns as early as possible as for non-key-patterns database scan is not required at all. General Terms Data Mining, Big Data, Data Science, Association Rule Mining,

Novel Frequent Pattern Mining Algorithm Based on Parallelization Scheme

Frequent pattern mining (FPM) is a very important technique in data mining and has attracted a wide range of practical applications. Equivalent Class Clustering (Eclat) has been identified as one of the most efficient FPM algorithm. We present P-Eclat, a novel parallel FPM algorithm which is an improvement of the Eclat algorithm, where a partial breadth-first search is employed to achieve maximum parallelism. Our approach uses a TIDset representation of the vertical transaction lists across multiple threads on a CPU. Current parallelization techniques for mining frequent patterns don’t fully utilize benefits accrued from multi core shared memory machines. Our parallel mining approach reduces the synchronization requirements, maximizing independence of data and enhances scalability. We also introduce several optimization techniques to improve the algorithm’s performance. Experimental results show that P-Eclat algorithm outperforms both Eclat and dEclat algorithms.

Recognition Framework for Inferring Activities of Daily Living Based on Pattern Mining

Ambient assisted living applications are very much dependent on robust activity recognition frameworks, which allow these applications to provide services based on the contextual information that has been discovered. Existing frameworks have generally focused on the application of traditional classifiers and semantics reasoning to recognize activities. Nevertheless, being able to recognize unexpected actions remains a challenge. The work in this paper presents an approach that is able to recognize activities that have been conducted in an unordered manner. The recognition framework extends an existing approach that recognizes activities by exploiting the different levels of abstraction within an activity. A frequent pattern mining algorithm has been applied to the recognition framework in order to find patterns within the stream of captured events, which in turn increases the adaptive learning ability of the proposed recognition framework. This paper also presents experimental results that validate the recognition ability of the recognition framework. The motivation of this work is to be able to detect the functional decline among elderly people suffering from Alzheimer’s disease by recognizing their daily activities.

An Improved PrePost Algorithm for Frequent Pattern Mining with Hadoop on Cloud

Due to the advancement in internet technologies the volume of data is tremendously increasing day by day. The research is gaining importance in extracting valuable information from such huge amount of data. Many research works are done and various algorithms are proposed. The PrePost algorithm is one of well-known algorithms of frequent pattern mining. It is based on N-list data structure to mine frequent item-sets. But the performance of PrePost algorithm degrades when it comes to processing of large amount of data. Hadoop is very well known technique for processing such large amount of data. This paper proposes the Improved PrePost algorithm which combines the features of Hadoop in order to process large data efficiently. Efficiency of PrePost algorithm is enhanced by implementing compact PPC tree with the general tree method and finding frequent itemsets without generating candidate itemsets. An architecture of the Improved PrePost algorithm with public cloud is proposed. The results show that as dataset size is increased, the Improved PrePost algorithm gives 60% better performance.

Efficiently mining frequent itemsets in transactional databases

Discovering frequent itemsets is an essential task in association rules mining and it is considered to be computationally expensive. To find the frequent itemsets, the algorithm of frequent pattern growth (FP-growth) is one of the best algorithms for mining frequent patterns. However, many experimental results have shown that building conditional FP-trees during mining data using this FP-growth method will consume most of CPU time. In addition, it requires a lot of space to save the FP-trees. This paper presents a new approach for mining frequent item sets from a transactional database without building the conditional FP-trees. Thus, lots of computing time and memory space can be saved. Experimental results indicate that our method can reduce lots of running time and memory usage based on the datasets obtained from the FIMI repository website.

Healthcare Analysis using Mining Algorithm on Mobile Technology

The healthcare analysis for the patient through the consulting service using mobile technology is proposed in this paper. A new MMM (Min Mapping Miner) process mining algorithm is introduced to overcome the limitation of existing pitfalls in H Mine algorithm. This algorithm has a unique feature. It uses limited memory and can work in parallel process. The analysis of healthcare handled with MMM, is actually a frequent pattern mining algorithm. The MMM frequent pattern mining algorithm has better confidence and performance for different treatment process for patients, where the doctors can envisage combination of treatments in the consulting service. Till now, the patient treatment has been represented using text or graphical representation, where as the MMM algorithm can represent using association rules. The confidences of association rules are exact to comprehend the patient treatment diagnosing process. The major focus of this article is to deal with the flow of treatment process in mobile device. A mobile technology is used to give the prompt understanding of the treatment given by the doctors to the patients, which are recorded in the form of event logs and stored as historical data. Hence, this development primarily contributes to knowledge discovery of various healthcare treatment diagnoses for the patient consulting services.

Occupancy-Based Frequent Pattern Mining *

Frequent pattern mining is an important data mining problem with many broad applications. Most studies in this field use support (frequency) to measure the popularity of a pattern, namely the fraction of transactions or sequences that include the pattern in a data set. In this study, we introduce a new interesting measure, namely occupancy, to measure the completeness of a pattern in its supporting transactions or sequences. This is motivated by some real-world pattern recommendation applications in which an interesting pattern should not only be frequent, but also occupies a large portion of its supporting transactions or sequences. With the definition of occupancy we call a pattern dominant if its occupancy value is above a user-specified threshold. Then, our task is to identify the qualified patterns which are both dominant and frequent. Also, we formulate the problem of mining top-k qualified patterns , that is, finding k qualified patterns with maximum values on a user-defined function of support and occupancy, for example, weighted sum of support and occupancy. The challenge to these tasks is that the value of occupancy does not change monotonically when more items are appended to a given pattern. Therefore, we propose a general algorithm called DOFRA (DOminant and FRequent pattern mining Algorithm) for mining these qualified patterns, which explores the upper bound properties on occupancy to drastically reduce the search process. Finally, we show the effectiveness of DOFRA in two real-world applications and also demonstrate the efficiency of DOFRA on several real and large synthetic datasets.

A framework for real-time Twitter data analysis

Twitter is a popular social network which allows millions of users to share their opinions on what happens all over the world. In this work we present a system for real-time Twitter data analysis in order to follow popular events from the user’s perspective. The method we propose extends and improves the Soft Frequent Pattern Mining (SFPM) algorithm by overcoming its limitations in dealing with dynamic, real-time, detection scenarios. In particular, in order to obtain timely results, the stream of tweets is organized in dynamic windows whose size depends both on the volume of tweets and time. Since we aim to highlight the user’s point of view, the set of keywords used to query Twitter is progressively refined to include new relevant terms which reflect the emergence of new subtopics or new trends in the main topic. The real-time detection system has been evaluated during the 2014 FIFA World Cup and experimental results show the effectiveness of our solution.

ENGLISH

Frequent itemset (or frequent pattern) mining is a very important issue in the data mining field. Both syntactic simplicity and descriptive potential, are the key features of the itemset-based pattern which have led to its widespread use in a growing number of real-life domains. Many efficient algorithms like Apriori & FP-Growth Algorithm are used to find frequent itemsets from large database. In almost all Frequent Pattern mining algorithms generating Frequent 1-itemsets are generated in order to find the support count(occurences) of each item in the entire transactions. This task is itself a tedious task in generating Frequent Patterns when considering the hugeness of modern databases available. No explicit strategy has been outlined in these algorithms to perform the aforesaid task. In this paper an efficient tree called Support Count tree has been proposed to perform this task. This algorithm can be easily embedded into any of the existing algorithms aimed at frequent pattern mining. With the help of this tree Frequent 1- Itemsets are found out quickly and efficiently which in-turn speeds up the generation of Frequent Patterns of the entire database.