Periodic Pattern Mining Research Articles

Stable top-k periodic high-utility patterns mining over multi-sequence

Stable top-k periodic high-utility patterns mining over multi-sequence

Mining fuzzy local periodic activity pattern for Smart home applications

Perceiving resident movements from ubiquitous sensor devices can aid health and safety management in smart home environment. One can find resident movement patterns to get notification about what should be aware of to prepare for possible accidents. The goal of this work is to propose a method to search for periodic residents’ activity patterns given smart home applications. Specifically, the proposed method is aimed to capture fuzzy local periodic activity patterns (FLPAP) from the smart home data. A FLPAP is an activity pattern that periodically occurs for a particular duration in fuzzy time intervals. We construct an algorithm to help to find the FLPAP called Fuzzy Local Periodic Activity Pattern Miner (FLoPMiner). On the side, a parameter called k-scale period is well-chosen to reorganize the data to make it suitable for the need of periodic activity pattern mining. The proposed method was demonstrated for its effectiveness on a public smart home dataset. After all, we successfully retrieved periodic resident activity patterns, which can be applied to self-health management systems, especially when focusing on a precise scale period. Overall, the proposed method shows descent pattern representation against the state-of-the-art methods of the same category.

Understanding Residents’ Behavior for Smart City Management by Sequential and Periodic Pattern Mining

Understanding the residents’ routine and repetitive behavior patterns is important for city planners and strategic partners to enact appropriate city management policies. However, the existing approaches reported in smart city management areas often rely on clustering or machine learning, which are ineffective in capturing such behavioral patterns. Aiming to address this research gap, this article proposes an analytical framework, adopting sequential and periodic pattern mining techniques, to effectively discover residents’ routine behavior patterns. The effectiveness of the proposed framework is demonstrated in a case study of American public behavior based on a large-scale venue check-in dataset. The dataset was collected in 2020 (during the global pandemic due to COVID-19) and contains 257 561 check-in data of 3995 residents. The findings uncovered interesting behavioral patterns and venue visit information of residents in the United States during the pandemic, which could help the public and crisis management in cities.

Mining motif periodic frequent travel patterns of individual metro passengers considering uncertain disturbances

Periodic pattern mining is of great significance for understanding passenger travel behavior, but the previous works mainly focused on the trajectory data and the dimension of the spot/point. Besides, many uncertain factors (severe weather, traffic accident, etc.) may interfere with discovering original and accurate periodic travel patterns. This paper proposes a novel type of travel pattern called motif periodic frequent pattern (MPFP), which captures the periodicity of network temporal motifs of individual metro passengers with higher-order spatio-temporal characteristics, considering, uncertain disturbances. We also propose a new complete mining algorithm MPFP-growth to extract MPFP from smart card data (SCD), and apply the real long-time-span experimental data from a large-scale metro system is applied. Results show that frequent-travel metro passengers usually have some typical MPFPs with the temporal periodic characteristic of “week”. Only the top 10 types of all 4 624 types account for about 95% of all motifs and the top 5 types constitute about 90%, and the MPFP of the top 3 types of motifs account for nearly 80% of all periodic patterns, in which Mono-MPFP and 2-MPFP are the main ones. The relatively stable time range of MPFP is three months, and the threshold for the optimal uncertain disturbance factor should be set at 5%. Additionally, several interesting typical MPFPs of individual metro commuting passengers and their proportions are introduced to further understand the multifarious variants of MPFP.

Comparison of Periodic Pattern Mining Algorithms on Temporal Datasets

Comparison of Periodic Pattern Mining Algorithms on Temporal Datasets

An advanced approach for incremental flexible periodic pattern mining on time-series data

Periodic pattern mining is a topic for mining periodic event patterns with sufficient confidence. The resulted patterns are often used to predict future events because they have established confidence that their periodicity will be maintained after the last time of the given data. Therefore, intelligent decision-making through periodic patterns has led to popularity in various applications, such as oil price fluctuations prediction, traffic congestion prediction, human behavior analysis, and sensor-based data analysis in the industry. Meanwhile, previous approaches adopting data structures based on the suffix-tree or trie developed the mining performance of this topic, and a concept that flexible periodic patterns consider don’t-care events in their intermediate events increased the flexibility of the results in this field. However, the data structures have limitations in terms of computing performance. Meanwhile, technological development has made collecting and accumulating data faster and faster. To conduct mining on modern databases which keep increasing incrementally, processing data streams in real time should be considered in the mining method. In this paper, we propose a novel flexible periodic pattern mining approach that can operate on incremental time-series databases by utilizing an enhanced data structure. Moreover, various experiments, such as multivariable analysis, sensitivity analysis, scalability analysis, demonstrate that the proposed approach has superior performance over previously proposed state-of-the-art algorithms in terms of runtime and memory usage.

Mining periodic high-utility itemsets with both positive and negative utilities

Mining high-utility patterns in databases containing items with both positive and negative profits is useful in market basket databases, since negative profits are common in the real world. Obviously, in the market basket database, patterns with stable long-term profits have more meaning. The discovery of itemsets with a consistent high frequency is known as periodic frequent pattern mining. Therefore, mining periodic high-utility patterns in a database containing items with both positive and negative profits is an interesting and useful task. However, this task has two main challenges. First, the utility measure does not have the download closure property. Second, the huge search space needs to be pruned more effectively. In this paper, we propose a vertical data structure-based algorithm called PHMN to discover periodic high-utility patterns (PHUPs) or itemsets in a transaction database with both positive and negative utilities. To be more efficient, we propose a new upper bound to prune the search space and an improved algorithm to discover the PHUPs. Finally, experiments are conducted to verify the effectiveness and efficiency of algorithms.

High Utility Periodic Frequent Pattern Mining in Multiple Sequences

Periodic pattern mining has become a popular research subject in recent years; this approach involves the discovery of frequently recurring patterns in a transaction sequence. However, previous algorithms for periodic pattern mining have ignored the utility (profit, value) of patterns. Additionally, these algorithms only identify periodic patterns in a single sequence. However, identifying patterns of high utility that are common to a set of sequences is more valuable. In several fields, identifying high-utility periodic frequent patterns in multiple sequences is important. In this study, an efficient algorithm called MHUPFPS was proposed to identify such patterns. To address existing problems, three new measures are defined: the utility, high support, and high-utility period sequence ratios. Further, a new upper bound, , and two new pruning properties were proposed. MHUPFPS uses a newly defined HUPFPS-list structure to significantly accelerate the reduction of the search space and improve the overall performance of the algorithm. Furthermore, the proposed algorithm is evaluated using several datasets. The experimental results indicate that the algorithm is accurate and effective in filtering several non-high-utility periodic frequent patterns.

Efficient mining of intra-periodic frequent sequences

Frequent Sequence Mining (FSM) is a fundamental task in data mining. Although FSM algorithms extract frequent patterns, they cannot discover patterns that periodically appear in the data. However, periodic trends are found in many areas such as market basket analysis, where discovering itemsets periodically purchased by customers can help understand periodic customer behavior. This is the task of Periodic Frequent Pattern Mining (PFPM). A major limitation common to traditional PFPM algorithms is that they reduce the periodicity between non-disjoint itemsets. They do not take into account the periods between disjoint itemsets. Thus, they find itemsets that appear periodically, but would fail to find a periodic appearance of distinct itemsets. To address this limitation, this paper extends the traditional problem of FSM with intra-periodicity and provides a theoretical background to extract intra-periodic frequent sequences. This leads to a new mining algorithm called Intra-Periodic Frequent Sequence Miner. Experimental results confirm its efficiency.

Fuzzy-driven periodic frequent pattern mining

Frequent pattern mining (FM) has a wide range of applications in the real world. But FM sometimes discovers many uninteresting patterns at the same time. Constraint-based FM, especially periodic constraint FM, becomes of interest, and it reduces redundant patterns. Most of the state-of-the-art algorithms for mining periodic frequent patterns (PFPs) are designed to find PFPs only in binary temporal databases. However, vast information in real situations is more suitable to be modelized as quantitative databases, in which most existing methods are inapplicable. Patterns obtained by those methods may change due to a tiny perturbation in the database, which is not expected in many situations. In this paper, we propose an algorithm (FP2M) to mine fuzzy-driven periodic frequent patterns and an algorithm (SFP2M) to find stable fuzzy-driven periodic frequent patterns in quantitative temporal databases. Fuzzy sets are used in our algorithms to deal with quantitative data since they provide a relaxed interval segmentation and are easy to implement. Novel pruning strategies and a new structure called Estimated Period Co-occurrence Structure (EPCS) are designed to speed up the mining process and improve efficiency. An improved method (SFP2M) is also presented by using the lability measurement to find stable patterns. Experimental evaluations on both real and synthetic datasets show good performance of the designed algorithms.

Mining relevant partial periodic pattern of multi-source time series data

Traditional partial periodic pattern mining algorithms tend to work on a single time series or database. However, time series databases usually consist of interrelated multivariate time series from different sources in real application scenarios. A new pattern–relevant partial periodic patterns and its corresponding mining algorithm (PMMS-Eclat) are designed, which can effectively reflect and mine the correlations of multi-source time series data. And PMMS-Eclat introduces the contribution difference of various patterns in the mining process to quickly identify important patterns. Firstly, PMMS-Eclat, an improved mining algorithm based on Ecalt, is used to obtain frequent partial periodic patterns from single time serie, which avoids the expensive tree structure adjustment cost. In this process, two new metrics are defined to characterize the correlation among sequential data effectively, namely density ratio and average periodic rate, which respectively reflect the density and occurrence rate of patterns. Secondly, the frequent partial periodic patterns excavated above are used as input matrices, and the ultimate relevant partial periodic patterns are obtained based on LSH principle. Finally, experiments on various datasets are conducted to verify the effectiveness of our proposed algorithm, and the results show that PMMS-Eclat is more competent for sparse databases.

An Efficient Algorithm for Mining Stable Periodic High-Utility Sequential Patterns

Periodic high-utility sequential pattern mining (PHUSPM) is used to extract periodically occurring high-utility sequential patterns (HUSPs) from a quantitative sequence database according to a user-specified minimum utility threshold (minutil). A sequential pattern’s periodicity is determined by measuring when the frequency of its periods (the time between two consecutive happenings of the sequential pattern) exceed a user-specified maximum periodicity threshold (maxPer). However, due to the strict judgment threshold, the traditional PHUSPM method has the problem that some useful sequential patterns are discarded and the periodic values of some sequential patterns fluctuate greatly (i.e., are unstable). In frequent itemset mining (FIM), some researchers put forward some strategies to solve these problems. Because of the symmetry of frequent itemset pattern (FIPs), these strategies cannot be directly applied to PHUSPM. In order to address these issues, this work proposes the stable periodic high-utility sequential pattern mining (SPHUSPM) algorithm. The contributions made by this paper are as follows. First, we introduce the concept of stability to overcome the abovementioned problems, mine sequential patterns with stable periodic behavior, and propose the concept of stable periodic high-utility sequential patterns (SPHUSPs) for the first time. Secondly, we design a new data structure named the PUL-list to record the periodic information of sequential patterns, thereby improving the mining efficiency. Thirdly, we propose the maximum lability pruning strategy in sequential pattern (MLPS), which can prune a large number of unstable sequential patterns in advance. To assess the algorithm’s effectiveness, we perform many experiments. It turns out that the algorithm can not only mine patterns that are ignored by traditional algorithms, but also ensure that the discovered patterns have stable periodic behavior. In addition, after using the MLPS pruning strategy, the algorithm can prune 46.5% of candidates in advance on average in six datasets. Pruning a large number of candidates in advance not only speeds up the mining process, but also greatly reduces memory usage.

Efficient Discovery of Partial Periodic Patterns in Large Temporal Databases

Periodic pattern mining is an emerging technique for knowledge discovery. Most previous approaches have aimed to find only those patterns that exhibit full (or perfect) periodic behavior in databases. Consequently, the existing approaches miss interesting patterns that exhibit partial periodic behavior in a database. With this motivation, this paper proposes a novel model for finding partial periodic patterns that may exist in temporal databases. An efficient pattern-growth algorithm, called Partial Periodic Pattern-growth (3P-growth), is also presented, which can effectively find all desired patterns within a database. Substantial experiments on both real-world and synthetic databases showed that our algorithm is not only efficient in terms of memory and runtime, but is also highly scalable. Finally, the effectiveness of our patterns is demonstrated using two case studies. In the first case study, our model was employed to identify the highly polluted areas in Japan. In the second case study, our model was employed to identify the road segments on which people regularly face traffic congestion.

SEQUENTIAL PATTERN MINING ALGORITHMS – RECENT TRENDS

Sequential pattern mining is a technique of data mining whose objective is to identify statistically relevant patterns within a database with time-related data. It has a wide range of applications in variety of domains like education, healthcare, bioinformatics, web usage mining, telecommunications, intrusion detection etc. At present, most of the real sequence databases are incremental in nature. So there is a need to explore incremental and distributed pattern mining algorithms. Periodic pattern mining is a technique to discover periodic pattern which may be a pattern that repeats itself after a specific time interval. It has a wide range of applications in weather prediction, stock market analysis, web usage recommendation etc. Moreover, uncertain frequent pattern mining has become a popular research domain among researchers, as many real-life databases at present consist of uncertain and incomplete data. In this paper, a novel attempt is made to incorporate a systematic literature review of state-of-the-art techniques of sequential pattern mining which ranges from incremental pattern mining, periodic pattern mining and uncertain frequent pattern mining. Researchers in the field of pattern mining will find it very useful to get the information about various algorithms of different types of pattern mining.

Mining Weighted Periodic Patterns by a Weighted Direction Graph Based Approach for Time-Series Databases

Periodic pattern mining in time series database plays an important part in data mining. However, most existing algorithms consider only the count of each item, but do not consider about the value of each item. To consider the value of each item on periodic pattern mining in time series databases, Chanda et al. proposed an algorithm called WPPM. In their algorithm, they construct the suffix trie to store the candidate pattern at first. However, the suffix trie would use too much storage space. In order to decrease the processing time for constructing the data structure, in this paper, we propose two data structures to store the candidates. The first data structure is Weighted Paired Matrix. After scanning the database, we will transform the database into the matrix type, and it is used for the second data structures. Therefore, our algorithm not only can decrease the usage of the memory space, but also the processing time. Because we do not need to use so much time to construct so many nodes and edges. Moreover, wealso consider the case of incremental mining for the increase of the data length. From the performance study, we show that our proposed algorithm based on the Weighted Direction Graphis more efficient than the WPPMalgorithm.

Identification of adverse disease agents and risk analysis using frequent pattern mining

Life-threatening illnesses such as cancer, cirrhosis of the liver, and hepatitis have become crucial problems for humanity. The risk of mortality can be deflated by early detection of symptoms and providing the best possible diagnosis. This critical role of detection and/or diagnosis can be enhanced using one of the techniques used in data mining, such as periodic pattern mining, association rule mining, classification. Analyzing the commonly occurring possible patterns or signs followed by performing the correlation analysis among those patterns can be exhaustively practiced for early detection and improve the diagnosis. Towards the adoption of association rule mining, devising a cost-effective and time-saving algorithm for mining frequent patterns plays an important role. In this paper, we propose an approach to pattern mining called Improved Frequent Pattern Growth (Improved FP-Growth). Firstly, it constructs an improvised frequent pattern tree data structure called Improved FP-tree. Moreover, Improved FP-Growth introduces a construction of conditional FP-tree data structure layout called Improved Conditional Frequent Pattern Tree (Improved Conditional FP-Tree). Unlike the traditional FP-Growth method, it uses both top-down and bottom-up approaches to efficiently generate frequent patterns without recursively constructing the improved conditional FP-tree. The experimental results emphasize the significance of the proposed Improved FP-Growth algorithm over a few traditional frequent itemset mining algorithms those adopt the approach of recursive conditional FP-tree construction.

Mining full, inner and tail periodic patterns with perfect, imperfect and asynchronous periodicity simultaneously

Periodic pattern has been utilized in many real life applications, such as weather conditions in a particular season, transactions in a superstore, power consumption, computer network fault analysis, and analysis of DNA and protein sequences. Periodic pattern mining is a popular though challenging research field in data mining because periodic patterns are of different types (namely full, inner, and tail patterns) and varied periodicities (namely perfect, imperfect, and asynchronous periodicity). Previous periodic pattern mining methods have some disadvantages: (1) Previous methods have to find different patterns separately; (2) They require postprocessing such as level-by-level join strategies for mining complex periodic patterns which have wildcards between two items. They cannot mine full, tail, and inner periodic patterns with perfect, imperfect, and asynchronous periodicities simultaneously. Therefore, an effective and comprehensive approach capable of discovering the above specified kinds of periodic patterns is needed. We propose a novel suffix tree-based algorithm, Mining dIfferent kinds of Periodic Patterns Simultaneously, MIPPS, to address the above issues. MIPPS finds different kinds of periodic patterns with different periodicities simultaneously without level-by-level join techniques using a novel incremental propagation generator. In addition, MIPPS mines periodic patterns efficiently using some pruning strategies. For the performance evaluation, we use both synthetic and real data to confirm good performance and scalability with complex periodic patterns.

TSPIN: mining top-k stable periodic patterns

Discovering periodic patterns consists of identifying all sets of items (values) that periodically co-occur in a discrete sequence. Although traditional periodic pattern mining algorithms have multiple applications, they have two key limitations. First, they consider that a pattern is not periodic if the time difference between two of its successive occurrences is greater than a maxPer threshold. But this constraint is too strict, as a pattern may be discarded based on only two of its occurrences, although it may be usually periodic. Second, traditional algorithms use a constraint that the support (occurrence frequency) of a pattern must be no less than a minSup threshold. But setting that parameter is not intuitive. Hence, it is usually set by trial and error, which is time-consuming. This paper addresses the first limitation by introducing a concept of stability to find periodic patterns that have a stable periodic behavior. Then, the second limitation is addressed by proposing an algorithm named TSPIN (Top-k Stable Periodic pattern mINer) to find the top-k stable periodic patterns, where the user can directly specify the number of patterns k to be found rather than using the minSup threshold. Several experiments have been performed to assess TSPIN’s performance, and it was found that it is efficient and can discover patterns that reveal interesting insights in real data.

HOVA-FPPM: Flexible Periodic Pattern Mining in Time Series Databases Using Hashed Occurrence Vectors and Apriori Approach

Finding flexible periodic patterns in a time series database is nontrivial due to irregular occurrence of unimportant events, which makes it intractable or computationally intensive for large datasets. There exist various solutions based on Apriori, projection, tree, and other techniques to mine these patterns. However, the existence of constant size tree structure, i.e., suffix tree, with extra information in memory throughout the mining process, redundant and invalid pattern generation, limited types of mined flexible periodic patterns, and repeated traversal over tree data structure for pattern discovery, results in unacceptable space and time complexity. In order to overcome these issues, we introduce an efficient approach called HOVA-FPPM based on Apriori approach with hashed occurrence vectors to find all types of flexible periodic patterns. We do not rely on complex tree structure rather manage necessary information in a hash table for efficient lookup during the mining process. We measured the performance of our proposed approach and compared the results with the baseline approach, i.e., FPPM. The results show that our approach requires lesser time and space, regardless of the data size or period value.

Periodic pattern mining from spatio-temporal database using novel global pollination artificial fish swarm optimizer-based clustering and modified FP tree

Periodic pattern mining from spatio-temporal database using novel global pollination artificial fish swarm optimizer-based clustering and modified FP tree