Sequential Pattern Mining Algorithm Research Articles

A personalized POI route recommendation system based on heterogeneous tourism data and sequential pattern mining

Planning a personalized POI route before touring a new city is an important travel preparation activity; however, it is a challenging and time-consuming task for tourists. Although some previous works focus on suggesting POI visit list or sequences, they fail to suggest personalized POI routes due to ignoring multifaceted tourism contexts. Also, they often suffer from tourist cold start or data sparsity problem because of the lack of tourism related data. To address the above weaknesses, we first propose a novel method to integrate heterogeneous tourism data collected from websites to construct a POI knowledgebase and massive structured POI visit sequences. Next, a POI-Visit sequential pattern mining algorithm is proposed to generate various fine-grained candidate POI routes from POI visit sequences while considering various tourism contexts. At the POI route recommendation stage, our system retrieve and rank a list of candidate routes according to the querying tourist’s tourism contexts, including the intended travel duration, the companion type in trip, the visit season and the preferring POI tourism types, etc. In our validation experiments, we select Guilin city in China as an example to construct a real POI knowledgebase which consists of 132 POIs and 8778 POI traffic time, and construct 5694 structured POI visit sequences based on 10,109 downloaded original travelogues. The experimental results demonstrate the advantages of our system in recommending fine-grained and high personalized POI routes for specific tourists.

Constraint-Based Sequential Pattern Mining with Decision Diagrams

Constraint-based sequential pattern mining aims at identifying frequent patterns on a sequential database of items while observing constraints defined over the item attributes. We introduce novel techniques for constraint-based sequential pattern mining that rely on a multi-valued decision diagram (MDD) representation of the database. Specifically, our representation can accommodate multiple item attributes and various constraint types, including a number of non-monotone constraints. To evaluate the applicability of our approach, we develop an MDD-based prefix-projection algorithm and compare its performance against a typical generate-and-check variant, as well as a state-of-the-art constraint-based sequential pattern mining algorithm. Results show that our approach is competitive with or superior to these other methods in terms of scalability and efficiency.

CNN-Based Fault Localization Method Using Memory-Updated Patterns for Integration Test in an HiL Environment

Automotive electronic components are tested via hardware-in-the-loop (HiL) testing at the unit and integration test stages, according to ISO 26262. It is difficult to obtain debugging information from the HiL test because the simulator runs a black-box test automatically, depending on the scenario in the test script. At this time, debugging information can be obtained in HiL tests, using memory-updated information, without the source code or the debugging tool. However, this method does not know when the fault occurred, and it is difficult to select the starting point of debugging if the execution flow of the software is not known. In this paper, we propose a fault-localization method using a pattern in which each memory address is updated in the HiL test. Via a sequential pattern-mining algorithm in the memory-updated information of the transferred unit tests, memory-updated patterns are extracted, and the system learns using a convolutional neural network. Applying the learned pattern in the memory-updated information of the integration test can determine the fault point from the normal pattern. The point of departure from the normal pattern is highlighted as a fault-occurrence time, and updated addresses are presented as fault candidates. We applied the proposed method to an HiL test of an OSEK/VDX-based electronic control unit. Through fault-injection testing, we could find the cause of faults by checking the average memory address of 3.28%, and we could present the point of fault occurrence with an average accuracy of 80%.

Exploring sequences of learner activities in relation to self-regulated learning in a massive open online course

Self-regulated learning (SRL) refers to how learners steer their own learning. Supporting SRL has been shown to enhance the use of SRL strategies and learning performance in computer-based learning environments. However, little is known about supporting SRL in Massive Open Online Courses (MOOCs). In this study, weekly SRL prompts were embedded as videos in a MOOC. We employed a sequential pattern mining algorithm, Sequential Pattern Discovery using Equivalence classes (cSPADE), on gathered log data to explore whether differences exist between learners who viewed the SRL-prompt videos and those who did not. Results showed that SRL-prompt viewers interacted with more course activities and completed these activities in a more similar sequential pattern than non SRL-prompt viewers. Also, SRL-prompt viewers tended to follow the course structure, which has been identified as a behavioral characteristic of students who scored higher on SRL (i.e., comprehensive learners) in previous research. Based on the results, implications for supporting SRL in MOOCs are discussed.

A MapReduce solution for incremental mining of sequential patterns from big data

Sequential Pattern Mining (SPM) is a popular data mining task with broad applications. With the advent of big data, traditional SPM algorithms are not scalable. Hence, many of the researchers have migrated to big data frameworks such as MapReduce and proposed distributed algorithms. However, the existing MapReduce algorithms assume the data as static and do not handle the incremental database updates. Moreover, they use to re-mine the updated database while new sequences are inserted. In this paper, we propose an efficient distributed algorithm for incremental sequential pattern mining (MR-INCSPM) using the MapReduce framework that can handle big data. The proposed algorithm incorporates the backward mining approach that efficiently makes use of the knowledge obtained during the previous mining process. Also, based on the study of item co-occurrences, we propose Co-occurrence Reverse Map (CRMAP) data structure. The issue of combinatorial explosion of candidate sequences is dealt using the proposed CRMAP data structure. Besides, a novel candidate generation and early prune mechanisms are designed using CRMAP to speed up the mining process. The proposed algorithm is evaluated on both the real and synthetic datasets. The experimental results prove the efficacy of MR-INCSPM with respect to processing time, memory and pruning efficiency.

Fast implementation of pattern mining algorithms with time stamp uncertainties and temporal constraints

Pattern mining is a powerful tool for analysing big datasets. Temporal datasets include time as an additional parameter. This leads to complexity in algorithmic formulation, and it can be challenging to process such data quickly and efficiently. In addition, errors or uncertainty can exist in the timestamps of data, for example in manually recorded health data. Sometimes we wish to find patterns only within a certain temporal range. In some cases real-time processing and decision-making may be desirable. All these issues increase algorithmic complexity, processing times and storage requirements. In addition, it may not be possible to store or process confidential data on public clusters or the cloud that can be accessed by many people. Hence it is desirable to optimise algorithms for standalone systems. In this paper we present an integrated approach which can be used to write efficient codes for pattern mining problems. The approach includes: (1) cleaning datasets with removal of infrequent events, (2) presenting a new scheme for time-series data storage, (3) exploiting the presence of prior information about a dataset when available, (4) utilising vectorisation and multicore parallelisation. We present two new algorithms, FARPAM (FAst Robust PAttern Mining) and FARPAMp (FARPAM with prior information about prior uncertainty, allowing faster searching). The algorithms are applicable to a wide range of temporal datasets. They implement a new formulation of the pattern searching function which reproduces and extends existing algorithms (such as SPAM and RobustSPAM), and allows for significantly faster calculation. The algorithms also include an option of temporal restrictions in patterns, which is available neither in SPAM nor in RobustSPAM. The searching algorithm is designed to be flexible for further possible extensions. The algorithms are coded in C++, and are highly optimised and parallelised for a modern standalone multicore workstation, thus avoiding security issues connected with transfers of confidential data onto clusters. FARPAM has been successfully tested on a publicly available weather dataset and on a confidential adult social care dataset, reproducing results obtained by previous algorithms in both cases. It has been profiled against the widely used SPAM algorithm (for sequential pattern mining) and RobustSPAM (developed for datasets with errors in time points). The algorithm outperforms SPAM by up to 20 times and RobustSPAM by up to 6000 times. In both cases the new algorithm has better scalability.

K/2-hop

With the increase of devices equipped with location sensors, mining spatio-temporal data for interesting behavioral patterns has gained attention in recent years. One of such well-known patterns is the convoy pattern which can be used, e.g., to find groups of people moving together in public transport or to prevent traffic jams. A convoy consists of at least m objects moving together for at least k consecutive time instants where m and k are user-defined parameters. Convoy mining is an expensive task and existing sequential algorithms do not scale to real-life dataset sizes. Existing sequential as well as parallel algorithms require a complex set of data-dependent parameters which are hard to set and tune. Therefore, in this paper, we propose a new fast exact sequential convoy pattern mining algorithm "k/2-hop" that is free of data-dependent parameters. The proposed algorithm processes the data corresponding to a few specific key timestamps at each step and quickly prunes objects with no possibility of forming a convoy. Thus, only a very small portion of the complete dataset is considered for mining convoys. Our experimental results show that k/2-hop outperforms existing sequential as well as parallel convoy pattern mining algorithms by orders of magnitude, and scales to larger datasets which existing algorithms fail on.

C3Ro: An efficient mining algorithm of extended-closed contiguous robust sequential patterns in noisy data

Sequential pattern mining has been the focus of many works, but still faces a tough challenge in the mining of large databases for both efficiency and apprehensibility of its resulting set. To overcome these issues, the most promising direction taken by the literature relies on the use of constraints, including the well-known closedness constraint. However, such a mining is not resistant to noise in data, a characteristic of most real-world data. The main research question raised in this paper is thus: how to efficiently mine an apprehensible set of sequential patterns from noisy data?In order to address this research question, we introduce 1) two original constraints designed for the mining of noisy data: the robustness and the extended-closedness constraints, 2) a generic pattern mining algorithm, C3Ro, designed to mine a wide range of sequential patterns, going from closed or maximal contiguous sequential patterns to closed or maximal regular sequential patterns. C3Ro is dedicated to practitioners and is able to manage their multiple constraints. C3Ro also is the first sequential pattern mining algorithm to be as generic and parameterizable. Extensive experiments have been conducted and reveal the high efficiency of C3Ro, especially in large datasets, over well-known algorithms from the literature. Additional experiments have been conducted on a real-world job offers noisy dataset, with the goal to mine activities. This experiment offers a more thorough insight into C3Ro algorithm: job market experts confirm that the constraints we introduced actually have a significant positive impact on the apprehensibility of the set of mined activities.

Campus Data Analysis Based on Positive and Negative Sequential Patterns

Campus data analysis is becoming increasingly important in mining students’ behavior. The consumption data of college students is an important part of the campus data, which can reflect the students’ behavior to a great degree. A few methods have been used to analyze students’ consumption data, such as classification, association rules, clustering, decision trees, time series, etc. However, they do not use the method of sequential patterns mining, which results in some important information missing. Moreover, they only consider the occurring (positive) events but do not consider the nonoccurring (negative) events, which may lead to some important information missing. So this paper uses a positive and negative sequential patterns mining algorithm, called NegI-NSP, to analyze the consumption data of students. Moreover, we associate students’ consumption data with their academic grades by adding the students’ academic grades into sequences to analyze the relationship between the students’ academic grades and their consumptions. The experimental results show that the students’ academic performance has significant correlation with the habits of having breakfast regularly.

CD-SPM: Cross-domain book recommendation using sequential pattern mining and rule mining

Recommender system suggests a personalized recommendation by filtering the information based on users interest. Nowadays, users like to purchase the best possible items and services to spend the shortest span of time. The cross-domain recommendation system is a method of recommendation wherein knowledge is gathered from multiple domains. With respect to the user’s search term from the source domain, most similar items are recommended from the target domain. Semantic similarity between two different items can be achieved through Wpath method using Ontology. PrefixSpan is used for generating sequential patterns and Topseq rule mining algorithm is used for finding the frequent sequential rule. So, this work tries to extend cross domain recommendation by 1) finding the semantic similarity of items using Ontology; 2) applying Collaborative Filtering for finding similar items and users; 3) generating frequent item sequences using PrefixSpan sequential pattern mining algorithm and 4) recommending user preferred items using Topseq rule mining algorithm. The recommender system is evaluated considering precision, recall and F1 Score measures. It finds CD-SPM which yields better F1 Score. The proposed approach also alleviates the new user problem and sparsity problem to some extent.

An efficient pixel clustering-based method for mining spatial sequential patterns from serial remote sensing images

The accumulation of serial remote sensing images provides plentiful data for discovering sequential spatial patterns in various fields such as agricultural monitoring, urban development, and vegetation cover. Otherwise, traditional sequential pattern-mining algorithms cannot be directly or efficiently applied to remote sensing images. In this study, we propose a pixel clustering-based method to improve the efficiency of mining spatial sequential patterns from raster serial remote sensing images (SRSI). Firstly, the images are compressed by using the Run-Length coding schema. Then, pixels with identical sequences are clustered by means of the Run-length code-based spatial overlay operation. Finally, a pruning strategy is proposed, to extend the prefixSpan algorithm to skip unnecessary database scanning when mining from pixel groups. The experimental results indicate that the method presented in this paper could extract spatial sequential patterns from SRSI efficiently. Although accurate support rates for the patterns may not be obtained, our method could ensure that all patterns are extracted with a lower time cost.

Location-Based Parallel Sequential Pattern Mining Algorithm

Given a data sequence, sequential pattern mining, which finds frequent sequence patterns among them, is an important data mining problem. However, in the existing sequential pattern mining, only the purchase order of the items is considered, and the position where the item is purchased is not considered. In this paper, we developed a sequential pattern mining algorithm using Apache spark. The proposed algorithm finds frequent sequential patterns in parallel by distributing data to several machines. Experimentally, we performed a comprehensive performance study on the proposed algorithm by varying various parameter values using various synthetic data. Experimental results show that the proposed algorithm shows a linear speed improvement over the number of machines.

A new Deep learning model for Sequential Pattern Mining using Ensemble learning and Models selection Taking Mobile Activity Recognition as a case

Sequential pattern mining has gained popularity in Data Mining and pattern recognition. Most sequential pattern mining algorithms are influenced by noisy variables, parameters tuning, bias-variance dilemma and learning instability. This paper presents a new deep learning model for sequential pattern mining, by using ensemble learning and models selection. Experimental studies on mobile activity recognition showed that our deep learning model, which is named Deep Sequential Pattern Mining (abbreviated as DeepSPM), obtained an enhanced generalization in comparison with Long Short Term Memory (LSTM), Bidirectional Associative Memory (BAM) and Hopfield. We provide a comparative performance analysis of pattern recognition. The advantages and the drawbacks of the benchmarking models are critically discussed.

Sequential Pattern Mining Suggests Wellbeing Supportive Behaviors

Amidst the headlines about the attention economy and the possible impacts of screen time, research investigating the complex relationship between digital technology usage and wellbeing has gained urgency. Researchers generally use a combination of surveys and automatic tracking tools to gather time and frequency of technology use. However, the focus of data analysis has been on measuring duration and frequency of usage rather than exploring behavioral patterns, possibly better indicators of mood states or stress levels. We propose a methodology for detecting behavioural patterns from digital footprints using a sequence pattern mining algorithm, and using these as features for predicting mood. Results show that our method can be used to analyze the relationship between digital usage and mood, and predict the latter with an accuracy of 80%, significantly above the baseline (71.1%). This method provides another angle to investigate digital technology usage in wellbeing-related research.

An Efficient Incremental Mining Algorithm for Discovering Sequential Pattern in Wireless Sensor Network Environments.

Wireless sensor networks (WSNs) are an important type of network for sensing the environment and collecting information. It can be deployed in almost every type of environment in the real world, providing a reliable and low-cost solution for management. Huge amounts of data are produced from WSNs all the time, and it is significant to process and analyze data effectively to support intelligent decision and management. However, the new characteristics of sensor data, such as rapid growth and frequent updates, bring new challenges to the mining algorithms, especially given the time constraints for intelligent decision-making. In this work, an efficient incremental mining algorithm for discovering sequential pattern (novel incremental algorithm, NIA) is proposed, in order to enhance the efficiency of the whole mining process. First, a reasoned proof is given to demonstrate how to update the frequent sequences incrementally, and the mining space is greatly narrowed based on the proof. Second, an improvement is made on PrefixSpan, which is a classic sequential pattern mining algorithm with a high-complexity recursive process. The improved algorithm, named PrefixSpan+, utilizes a mapping structure to extend the prefixes to sequential patterns, making the mining step more efficient. Third, a fast support number-counting algorithm is presented to choose frequent sequences from the potential frequent sequences. A reticular tree is constructed to store all the potential frequent sequences according to subordinate relations between them, and then the support degree can be efficiently calculated without scanning the original database repeatedly. NIA is compared with various kinds of mining algorithms via intensive experiments on the real monitoring datasets, benchmarking datasets and synthetic datasets from aspects including time cost, sensitivity of factors, and space cost. The results show that NIA performs better than the existed methods.

Sequential Pattern Mining Algorithm Based on Text Data: Taking the Fault Text Records as an Example

Sequential pattern mining (SPM) is an effective and important method for analyzing time series. This paper proposed a SPM algorithm to mine fault sequential patterns in text data. Because the structure of text data is poor and there are many different forms of text expression for the same concept, the traditional SPM algorithm cannot be directly applied to text data. The proposed algorithm is designed to solve this problem. First, this study measured the similarity of fault text data and classified similar faults into one class. Next, this paper proposed a new text similarity measurement model based on the word embedding distance. Compared with the classic text similarity measurement method, this model can achieve good results in short text classification. Then, on the basis of fault classification, this paper proposed the SPM algorithm with an event window, which is a time soft constraint for obtaining a certain number of sequential patterns according to needs. Finally, this study used the fault text records of a certain aircraft as experimental data for mining fault sequential patterns. Experiment showed that this algorithm can effectively mine sequential patterns in text data. The proposed algorithm can be widely applied to text time series data in many fields such as industry, business, finance and so on.

The concept of sequential pattern mining for text

Sequential pattern mining is one of popular data mining technique with sequential pattern as representation of data. However, most of sequential pattern mining research was conducted for structured data. In this paper, we did literature review of the sequential pattern mining algorithm that suitable for unstructured data such as text data. We reviewed several sequential pattern mining algorithm that had already used in text mining research, among others GSP, Spade, PrefixSpan, Spam, Lapin, SM-Spam, CM-Spade, BIDE, and another various algorithm based on sequential pattern mining problem such as concise representation and how to extract more rich pattern. The result showed that that from year to year research on text data using sequential pattern mining had increased. Although, not many algorithm were developed and also still rarely new algorithms were implemented in text data.

Preprocessing technology of consuming big data based on user interest with internet of location mining

ABSTRACT The current group discovery research mainly uses the communication relationship to mine user groups and fails to make full use of the user social relationships implied in the network, so that the mining groups can not truly reflect the user’s social relations in the social life. In this paper, the consumer big data preprocessing technology based on user interest location mining is proposed. The sequence pattern mining algorithm is used to mine user’s positional regularity sequence, and the position sequence correlation metrics are established. The positional regularity correlation is used to reveal the social relationship of the user’s social activities. Second, this paper combines the local similarity measure method to calculate the user’s communication distance index, which reflects the degree of understanding between users. Finally, using the communication distance index to weight the position correlation to calculate the user group correlation and then use the split clustering algorithm to mine user groups with communication relationship and social relationship. The experimental results showed that this method can effectively mine the communication relevance and position correlation in the user social activities and reflect the user’s group relations in real social activities.

Discovering individual movement patterns from cell-id trajectory data by exploiting handoff features

The discovery of movement patterns from trajectory data is crucial for supporting many location-based applications. Most existing methods require that the trajectories contain explicit location information. However, it is usually difficult to collect such kind of trajectories from mobile phone users. In this paper, we propose a method for mining movement patterns from cell-id trajectories (i.e., sequences of cell tower identifiers) without explicit location information. Specifically, we firstly estimate the spatial closeness between cell towers in a cell-id trajectory dataset by exploiting the handoff features. Then, we propose a novel sequential pattern mining algorithm to mine movement patterns from the cell-id trajectory dataset by taking into account the estimated spatial closeness. We evaluated the proposed method based on a real cell-id trajectory dataset. The experiment results show that the proposed method can adapt to the high uncertainty of cell-id trajectories and it outperforms state-of-the-art methods in terms of efficiency, completeness, and usefulness.

Exploring multi-stage driving behaviours prior to potential vehicle–pedestrian collisions

A simulator-based experiment involving a vehicle–pedestrian collision scenario was designed and the corresponding temporal patterns of the multi-stage driving behaviours were then analysed using the approach of sequential pattern mining. The test scenario, conducted by 45 participants, involved a vehicle travelling along a straight two-lane roadway when an unexpected pedestrian emerges and runs across the driving lane. The whole process was partitioned into three stages – the pre-event stage, the evasive stage and the conflict stage – and several featured variables (braking reaction time, time-to-collision etc.) were collected. A sequential pattern mining algorithm was adopted to analyse the temporal relationships among these variables and, for comparison, several conventional binary regression models were also developed. The analysis revealed several sequential patterns of driving behaviours under the crash-avoidance scenario. For example, the drivers who experienced collisions were accelerating during the first stage and also had relatively long braking reaction times in the second stage. The results from the sequential pattern mining analysis indicate that the evasive stage is crucial for understanding the whole process, as it affects the influence of behaviours from the pre-event stage to the conflict stage.