20 June 2025 in IEEE Access
Partial periodic-frequent pattern mining is a critical technique in the data mining field. This technique finds all frequent patterns demonstrating partial periodicity within temporal datasets. Despite its potential, the broader industrial adoption of this technique has been limited due to two significant challenges. First, there is currently no algorithm capable of efficiently identifying the required patterns in columnar temporal databases, which are increasingly prevalent in real-world applications. Second, the existing algorithms suffer from high computational demands regarding processing time (or runtime) and memory usage, rendering them unsuitable for analyzing large-scale datasets effectively. To address these challenges, this paper proposes the Generalized Partial Periodic-Frequent Depth-First Search (GPPF-DFS) algorithm. GPPF-DFS compresses row-based and columnar temporal databases into a unified dictionary structure, enabling efficient data traversal and recursive mining. The experimental results across various datasets demonstrate that our proposed algorithm significantly outperforms GPF-growth in terms of effectiveness and computational efficiency. Furthermore, a scalability test has been conducted to assess the performance of the proposed algorithm. Additionally, we demonstrate the practical applicability of GPPF-DFS through case studies on air pollution and retail analytics. The algorithm identifies periodic patterns in air pollution data, offering valuable insights for environmental monitoring. In retail analytics, it uncovers patterns that inform business strategies, and enhance customer experience.
6 June 2025 in IEEE Transactions on Fuzzy Systems
Partial periodic patterns play a significant role in identifying regularities within temporal databases. However, most existing research has focused on discovering these patterns in binary datasets, overlooking the critical insights about the quantitative values associated with the items. This study utilizes the principles of fuzzy sets to propose a novel model for discovering Fuzzy Partial Periodic Frequent Patterns (FPPFPs) within a quantitative temporal database. A robust depth-first search algorithm has also been proposed to uncover all FPPFPs. The proposed algorithm incorporates a novel pruning strategy that effectively reduces the search space and the computational cost required for discovering the FPPFPs. The experimental findings on synthetic and real-world datasets demonstrate the efficiency of the proposed algorithm. Finally, a case study utilizing air pollution data is provided to showcase the practical applicability and significance of the identified patterns.
16 June 2025 in Scientific Data
Air pollution is a leading risk factor for mortality in Japan. Despite improvements in air quality, Particulate Matter 2.5 (PM2.5) remains a significant concern due to its severe health impacts. Although previous studies have analyzed PM2.5 trends in Japan, they often overlook regions consistently exposed to high PM2.5 levels, primarily due to the lack of comprehensive, high-quality datasets. To address this gap, we developed a large-scale dataset containing five years of hourly PM2.5 measurements from approximately 1,900 sensors across Japan. The dataset employs an ER-Schema that isolates PM2.5 data, ensuring a focused analysis by excluding other pollutants. This extensive temporal and spatial coverage enables a more precise examination of long-term PM2.5 exposure patterns. Finally, we validated the dataset by applying established pattern mining models, demonstrating its effectiveness in identifying regions with consistently high PM2.5 levels. The proposed dataset provides a valuable resource for air quality research, supporting the development of targeted mitigation strategies and enhancing understanding of pollution dynamics in Japan.
4 November 2024 in IEEE Access
Geo referenced time series is a cornerstone in spatiotemporal data analysis, revealing invaluable patterns that drive socio-economic development. Previous studies have modeled these series as transactional databases. However, traditional transactional databases fail to account for the sequential order of items in a series, leading to the incomplete identification of patterns in applications where the sequence order is significant. Motivated by this limitation, this paper introduces a novel data transformation technique that converts geo-referenced time series into a sequential database, preserving the order of items. This approach is advantageous over traditional models because it maintains the temporal sequence, which is crucial for accurate pattern identification. We present a novel model for discovering geo-referenced frequent sequential patterns (GFSPs) within this transformed database. These patterns are termed "interesting" due to their ability to uncover significant temporal and spatial relationships within the data, which are crucial for understanding complex phenomena like traffic congestion. Our method reveals patterns that are not only frequent but also hold practical implications for traffic management and urban planning. Additionally, we introduce a neighborhood-aware exploration technique designed to optimize search space and computational efficiency. Experimental results validate the effectiveness and efficiency of our proposed algorithm. The utility of the identified patterns is demonstrated through a case study focused on frequent congestion patterns in road network data, underscoring the relevance of the sequential database approach.
1 August 2026 in IEEE Transactions on Fuzzy Systems