Frequent Sequence Mining Research Articles

Exploring interactive behaviours of urban and rural teachers in blended synchronous classrooms: Insights from a proposed interaction analysis framework

Blended synchronous classrooms (BSCs) play a critical role in narrowing the educational gap between urban and rural areas in China, promoting educational equity. In BSCs, the quantity and quality of interactive teaching behaviours of urban and rural teachers may significantly influence the learning experiences of students in both urban and rural settings, and they are supposed to be different from interactive teaching behaviours that happen in traditional classrooms. However, there is limited research on the characteristics of such interactive teaching behaviours in BSCs. To address the gap, this study first proposed an interaction analysis framework for BSCs (IAF4BSCs) and then examined the interactive teaching behaviours of urban and rural teachers in BSCs using this framework, along with various learning analytics methods including lag sequential analysis, frequent sequence mining and thematic analysis. The data analysis yielded the following results: (1) Urban teachers exhibited more speech behaviours such as lecturing and providing immediate feedback, whereas rural teachers engaged more in non-speech behaviours like classroom patrols. (2) Despite the expectation for urban teachers to interact equally with urban and rural students in BSCs, they predominantly focused on urban students. (3) Urban teachers demonstrated diverse interaction sequences with both urban and rural students, whereas rural teachers primarily interacted with rural students. (4) Interviews with teachers and students revealed both positive and negative aspects of interactive teaching behaviours in BSCs. Based on these findings, the study proposed three pedagogical recommendations to enhance interactive teaching behaviours in BSCs: increasing interactive participation, integrating technology more effectively and emphasizing post-class assessment.

MFS-SubSC: an efficient algorithm for mining frequent sequences with sub-sequence constraint

MFS-SubSC: an efficient algorithm for mining frequent sequences with sub-sequence constraint

Privacy-Enhanced Frequent Sequence Mining and Retrieval for Personalized Behavior Prediction

Privacy-Enhanced Frequent Sequence Mining and Retrieval for Personalized Behavior Prediction

Making strides towards AI-supported regulation of learning in collaborative knowledge construction

Collaborative knowledge construction (CKC) requires individual group members to share information and resources, sustain the improvement of ideas through peer interactions, and the construction and development of collective knowledge at the group level. Investigations of this multi-level nature are of critical importance for developing AI-based regulation support systems for learners’ CKC. This paper proposes a framework using an integrated analytics approach (Hidden Markov Model combined with Lag sequential analysis and Frequent sequence mining) for analyzing multi-level characteristics of CKC at cognitive and regulative behaviour dimensions. The approach was applied to a process-oriented discourse dataset of groups working collaboratively on concept-mapping activities. The results showed that the suggested approach can reveal insights into the multi-level (i.e., regulation behaviours likely to start at the group level, move to the peer-level and occasionally to individual-level self-regulation) and the dynamic nature of CKC (i.e., CKC appears to start from a group-level regulation pattern to move to a cognitive behaviour pattern of perspective sharing at the individual and peer levels, going back to regulation at individual and peer-levels and most likely end with a cognitive behaviour pattern at the group level). The novel models presented here can be used to classify learner behaviours into CKC states, enabling personalized scaffolding opportunities based on detected patterns in future intelligent support systems.

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.

Using learning analytics to investigate learning processes and behavioural patterns in an augmented reality language learning environment

AbstractBackgroundAugmented reality has been widely applied in various fields, and its benefits in language learning have been increasingly recognized. However, the investigation of effective learning behaviours and processes in augmented reality learning environments, taking into account temporality and analysis of differences in learning behaviours between learning achievements from the perspective of learning strategies, is still a topic that requires further research.ObjectivesThe goal of this research is to explore effective learning processes in AR language learning environments and learning behavioural patterns of learners with different performance levels via LA taking into account temporality.MethodsWe used the learning application designed and developed in previous studies to understand the learning process and behavioural patterns of acquiring Japanese compound verbs with augmented reality. Lag sequential analyses were conducted to analyse the learning behavioural transitions using learning data. Additionally, frequent sequences were analysed using frequent sequence mining to compare the learning behavioural patterns of learners with different learning performances.Results and ConclusionsWe investigated the learning process and behavioural patterns and discussed the language learning strategies employed in the augmented reality learning environment. The results indicated that learners primarily performed strategies of performed expressions, recombining, and structured reviewing in AR compound verb learning activities. Furthermore, the results revealed the different types of strategies applied by different learning performance learners, whereby the low group learners used the least variety of strategies, and the middle and high group learners employed repeating and deductive reasoning strategies that were not followed by the low group.

IMCSA: Providing Better Sequence Alignment Space for Industrial Control Protocol Reverse Engineering

Nowadays, with the wide application of industrial control facilities, industrial control protocol reverse engineering has significant security implications. The reverse method of industrial protocol based on sequence alignment is the current mainstream method because of its high accuracy. However, this method will incur a huge time overhead due to unnecessary alignments during the sequence alignment process. In this paper, we optimize the traditional sequence alignment method by combining the characteristics of industrial control protocols. We improve the frequent sequence mining algorithm, Apriori, to propose a more efficient Bag-of-Words generation algorithm for finding keywords. Then, we precluster the messages based on the generated Bag-of-Words to improve the similarity of the message within a cluster. Finally, we propose an industrial control protocol message preclustering model for sequence alignment, namely, IMCSA. We evaluate it over five industrial control protocols, and the results show that IMCSA can generate clusters with higher message similarity, which will greatly reduce the invalid alignments existing in the sequence alignment stage and ultimately improve the overall efficiency.

Local differential privacy-based frequent sequence mining

Frequent sequence mining (FSM) is a fundamental component for analyzing sequential data in the big data era. However, collecting and analyzing sequence data incurs serious privacy issues for users. Local differential privacy (LDP) is a prevalent privacy paradigm for data statistics and analysis. Due to the high-dimensionality of sequence data, directly applying LDP on FSM faces severe challenges such as heavy perturbation, waste of privacy budget, poor accuracy, and high computational costs. This paper proposed PrivFSM and PrivFSM∗ which are two LDP-compliant FSM mechanisms with high data utility and low computational cost. Specifically, PrivFSM iteratively constructs sequences into a prefix tree under LDP to mine frequent sequences from huge domain. Besides, PrivFSM leverages a threshold-based pruning strategy to prune fake nodes, which reduces perturbations and computation costs. As an augmented version of PrivFSM, PrivFSM∗ no longer splits privacy budget, but partitions user set to avoid heavy noise. PrivFSM∗ maximizes the utilization of users’ data by applying the whole data on every iteration with a complete privacy budget. We conduct theoretical analysis on error-bound and extensive experiments on both real-world and synthetic datasets. The experimental results demonstrate the high data utility of our proposed algorithms when compared with other state-of-the-art mechanisms.

Mining Simple Path Traversal Patterns in Knowledge Graph

The data mining has remained a subject of unfailing charm for research. The knowledge graph is rising and showing infinite life force and strong developing potential in recent years, where it is observed that acyclic knowledge graph has capacity for enhancing usability. Though the development of knowledge graphs has provided an ample scope for appearing the abilities of data mining, related researches are still insufficient. In this paper, we introduce path traversal patterns mining to knowledge graph. We design a novel simple path traversal pattern mining framework for improving the representativeness of result. A divide-and-conquer approach of combining each path is proposed to discover the most frequent traversal patterns in knowledge graph. To support the algorithm, we design a linked list structure indexed by the length of sequences with handy operations. The correctness of algorithm is proven. Experiments show that our algorithm reaches a high coverage with low output amounts compared to existing frequent sequence mining algorithms.

Smart Multimedia Compressor—Intelligent Algorithms for Text and Image Compression

AbstractA number of text compression algorithms have been proposed in the past decades, which have been very effective and usually operate on conventional character/word-based approaches. A novel data compression perspective of data mining is explored in this research and the paper focuses on novel frequent sequence/pattern mining approach to text compression. This work attempts to make use of longer-range correlations between words in languages for achieving better text compression. We propose a novel and efficient method by making the compression of any word-level text in a universal manner for corpora across domains referred as Universal Huffman Tree-based encoding. The major contribution of this work is in terms of avoidance of code table communication to the decoder. Simulation results over benchmark datasets indicate that Universal Huffman encoding employing frequent sequence mining (achieves [20%, 89%] improvement in compression in reduced time. The paper also contributes a usable interface for Data compression that employs the proposed frequent sequent mining-based data compression algorithm. The interface supports features such as feedback, consistency, usability, navigation, visual appeal, performance and accessibility on par with existing compression softwares. The work results in an intelligent data compression software employing knowledge engineering perspective.

A Survey of Data Mining Activities in Distributed Systems

Distributed systems, which may be utilized to do computations, are being developed as a result of the fast growth of sharing resources. Data mining, which has a huge range of real applications, provides significant techniques for extracting meaningful and usable information from massive amounts of data. Traditional data mining methods, on the other hand, suppose that the data is gathered centrally, stored in memory, and is static. Managing massive amounts of data and processing them with limited resources is difficult. Large volumes of data, for instance, are swiftly generated and stored in many locations. This becomes increasingly costly to centralize them at a single location. Furthermore, traditional data mining methods typically have several issues and limitations, such as memory restrictions, limited processing ability, and insufficient hard drive space, among others. To overcome the following issues, distributed data mining's have emerged as a beneficial option in several applications According to several authors, this research provides a study of state-of-the-art distributed data mining methods, such as distributed common item-set mining, distributed frequent sequence mining, technical difficulties with distributed systems, distributed clustering, as well as privacy-protection distributed data mining. Furthermore, each work is evaluated and compared to the others.

Frequent high minimum average utility sequence mining with constraints in dynamic databases using efficient pruning strategies

Frequent high minimum average utility sequence mining with constraints in dynamic databases using efficient pruning strategies

Privacy Protection of Frequent Sequences based on Uniform Adaptive Local Margin

In order to enhance the privacy protection of frequent sequences, and improve its mining utility and reduce the impact of data dimension, this paper proposes a consistent adaptive margin for marginal release under local difference privacy, which has the characteristics of improving the effectiveness and efficiency. The algorithm mainly uses the consistent adaptive local margin strategy to index, finds frequent sequences, perturbs transaction data through random response mechanism, and finds all the sensitivities that satisfy local differential privacy. For the purpose of determining the running time of mining frequent sequences, a reasonable local sensitivity traversal data set before and after the interference is selected. According to the combination property of differential privacy, it is proved theoretically that the algorithm satisfies local differential privacy, and the effectiveness of the algorithm is verified by experiments. The experimental results show that the proposed algorithm can protect the local differential privacy of frequent sequences safely and efficiently, and ensure the accuracy of frequent sequences.

Efficient algorithms for mining frequent high utility sequences with constraints

An important data mining task is to discover all high utility sequences in a quantitative sequence database. Although useful, the number of discovered sequences is often very large. To find patterns that are more tailored to a user’s needs, this paper studies the problem of mining frequent high utility sequences satisfying item constraints. This article proposes a novel algorithm named C-FHUSM to quickly obtain these sequences from two concise representations discovered from a quantitative sequence database, namely frequent generator high utility sequences and frequent closed high utility sequences. The first set is extracted using a novel algorithm named FGenHUSM, while an existing algorithm is applied to extract the second set. C-FHUSM integrates novel pruning techniques to ignore sequences that do not satisfy item constraints early by checking only a small number of representative sequences at the beginning of the mining process. Experimental results show that C-FHUSM can be more than ten times faster and has better scalability than a modified version of the state-of-the-art EHUSM algorithm for mining sequences with item constraints. Moreover, it is found that using C-FHUSM is beneficial when a user frequently changes constraints as results can be updated without rescanning the database.

Protocol format extraction based on an improved CFSM algorithm

As the information technology rapidly develops, many network applications appear and their communication protocols are unknown. Although many protocol keyword recognition based protocol reverse engineering methods have been proposed, most of the keyword recognition algorithms are time consuming. This paper firstly uses the traffic clustering method F-DBSCAN to cluster the unknown protocol traffic. Then an improved CFSM (Closed Frequent Sequence Mining) algorithm is used to mine closed frequent sequences from the messages and identify protocol keywords. Finally, CFGM (Closed Frequent Group Mining) algorithm is proposed to explore the parallel, sequential and hierarchical relations between the protocol keywords and obtain accurate protocol message formats. Experimental results show that the proposed protocol formats extraction method is better than Apriori algorithm and Sequence alignment algorithm in terms of time complexity and it can achieve high keyword recognition accuracy. Additionally, based on the relations between the keywords, the method can obtain accurate protocol formats. Compared with the protocol formats obtained from the existing methods, our protocol format can better grasp the overall structure of target protocols and the results perform better in the application of protocol reverse engineering such as fuzzing test.

Mining Integrated Sequential Patterns From Multiple Databases

Existing work on multiple databases (MDBs) sequential pattern mining cannot mine frequent sequences to answer exact and historical queries from MDBs having different table structures. This article proposes the transaction id frequent sequence pattern (TidFSeq) algorithm to handle the difficult problem of mining frequent sequences from diverse MDBs. The TidFSeq algorithm transforms candidate 1-sequences to get transaction subsequences where candidate 1-sequences occurred as (1-sequence, itssubsequenceidlist) tuple or (1-sequence, position id list). Subsequent frequent i-sequences are computed using the counts of the sequence ids in each candidate i-sequence position id list tuples. An extended version of the general sequential pattern (GSP)-like candidate generates and a frequency count approach is used for computing supports of itemset (I-step) and separate (S-step) sequences without repeated database scans but with transaction ids. Generated patterns answer complex queries from MDBs. The TidFSeq algorithm has a faster processing time than existing algorithms.

Fast generation of sequential patterns with item constraints from concise representations

Constraint-based frequent sequence mining is an important and necessary task in data mining since it shows results very close to the requirements and interests of users. Most existing algorithms for performing this task are based on a traditional approach that mines patterns directly from a sequence database (SDB). However, in fact, SDBs are often very large. The algorithms thus often exhibit poor performance because the number of generated candidates and the search space are enormous, especially for low minimum support thresholds. In addition, these algorithms must read an SDB again when a constraint is changed by the user. In the context of frequently varied constraints, repeatedly scanning SDBs consume much time. To address this issue, we propose a novel approach for generating frequent sequences with various constraints from the two sets of frequent closed sequences ($$ {{\mathcal{F}}{\mathcal{C}}{\mathcal{S}}} $$) and frequent generator sequences ($$ {{\mathcal{F}}{\mathcal{G}}{\mathcal{S}}} $$), which are the concise representations of the set $$ {{\mathcal{F}}{\mathcal{S}}} $$ of all frequent sequences. The proposed approach is based on novel theoretical results that show an explicit relationship between $$ {{\mathcal{F}}{\mathcal{S}}} $$ and these two sets and have been strictly proved. The approach is then used to develop an efficient algorithm named MFS-IC for quickly generating frequent sequences with item constraints, a task that has many real-life applications. Extensive experiments on real-life and synthetic databases show that the proposed MFS-IC algorithm outperforms state-of-the-art algorithms, which directly mine frequent sequences with constraints from an SDB, in terms of runtime, memory usage and scalability.

Learning and predicting operation strategies by sequence mining and deep learning

The operators of chemical technologies are frequently faced with the problem of determining optimal interventions. Our aim is to develop data-driven models by exploring the consequential relationships in the alarm and event-log database of industrial systems. Our motivation is twofold: (1) to facilitate the work of the operators by predicting future events and (2) analyse how consequent the event series is. The core idea is that machine learning algorithms can learn sequences of events by exploring connected events in databases. First, frequent sequence mining applications are utilised to determine how the event sequences evolve during the operation. Second, a sequence-to-sequence deep learning model is proposed for their prediction. The long short-term memory unit-based model (LSTM) is capable of evaluating rare operation situations and their consequential events. The performance of this methodology is presented with regard to the analysis of the alarm and event-log database of an industrial delayed coker unit.

A Unified Framework for Frequent Sequence Mining with Subsequence Constraints

Frequent sequence mining methods often make use of constraints to control which subsequences should be mined. A variety of such subsequence constraints has been studied in the literature, including length, gap, span, regular-expression, and hierarchy constraints. In this article, we show that many subsequence constraints—including and beyond those considered in the literature—can be unified in a single framework. A unified treatment allows researchers to study jointly many types of subsequence constraints (instead of each one individually) and helps to improve usability of pattern mining systems for practitioners. In more detail, we propose a set of simple and intuitive “pattern expressions” to describe subsequence constraints and explore algorithms for efficiently mining frequent subsequences under such general constraints. Our algorithms translate pattern expressions to succinct finite-state transducers, which we use as computational model, and simulate these transducers in a way suitable for frequent sequence mining. Our experimental study on real-world datasets indicates that our algorithms—although more general—are efficient and, when used for sequence mining with prior constraints studied in literature, competitive to (and in some cases superior to) state-of-the-art specialized methods.

HUPSMT: AN EFFICIENT ALGORITHM FOR MINING HIGH UTILITY-PROBABILITY SEQUENCES IN UNCERTAIN DATABASES WITH MULTIPLE MINIMUM UTILITY THRESHOLDS

The problem of high utility sequence mining (HUSM) in quantitative se-quence databases (QSDBs) is more general than that of frequent sequence mining in se-quence databases. An important limitation of HUSM is that a user-predened minimum tility threshold is used commonly to decide if a sequence is high utility. However, this is not convincing in many real-life applications as sequences may have diferent importance. Another limitation of HUSM is that data in QSDBs are assumed to be precise. But in the real world, collected data such as by sensor maybe uncertain. Thus, this paper proposes a framework for mining high utility-probability sequences (HUPSs) in uncertain QSDBs (UQS-DBs) with multiple minimum utility thresholds using a minimum utility. Two new width and depth pruning strategies are also introduced to early eliminate low utility or low probability sequences as well as their extensions, and to reduce sets of candidate items for extensions during the mining process. Based on these strategies, a novel ecient algorithm named HUPSMT is designed for discovering HUPSs. Finally, an experimental study conducted in both real-life and synthetic UQSDBs shows the performance of HUPSMT in terms of time and memory consumption.