Article Preview
TopIntroduction
Organizations worldwide leverage massive amounts of data to add value in everyday development processes; internal data providers release data volumes specifically for use in development processes. In automotive engineering, this sensor data is essential for semantically understanding and identifying faults in vehicle components. The high degree of networking and complexity of individual components and systems in vehicles presents significant challenges (Dai & Gao, 2013). Advanced techniques are required, therefore, to facilitate efficient data provision and anomaly detection, especially given the rapid growth of datasets (Johanson et al., 2014). Automotive manufacturers rely increasingly on advanced data analysis methods to monitor the performance of various systems and components (L’Heureux et al., 2017).
As data initially exists in a raw state, researchers optimize their compression using big data methods. big data systems assist organizations in efficiently collecting, storing, and analyzing large datasets, providing insights into vehicle behavior and emerging trends in engines (L’Heureux et al., 2017). Through complex analytics, machine learning (ML), and semantic understanding based on taxonomies, companies can gain useful insights to enhance decision-making, develop customer-centric strategies, and personalize experiences of the domain experts (Carreno et al., 2020; Tichomirov et al., 2023). This taxonomy aims to provide a better semantic understanding of measurement data and events from the perspective of engine development. It contributes to the field of the semantic web—an extension of the current web—designed to make information understandable not only to humans but also to machines (Berners-Lee et al., 2023). The goal is to structure data in a way that allows algorithms to efficiently process, interpret, and utilize it. According to Tichomirov et al. (2023), this taxonomy enables ML algorithms to comprehend the meaning of data, moving beyond mere anomaly detection, enabling a deeper, context-sensitive analysis of measurement data and events relevant to engine development. Additionally, it also allows artificial intelligence systems to infer new knowledge from existing datasets. This ability to infer new insights uncovers hidden connections and trends, leading to innovative approaches and more efficient reasoning in engine development. Hence, in specific fields like engine development, robust big data platforms are crucial for managing the information flood and optimizing customer-centric activities (L’Heureux et al., 2017; Ullah et al., 2024; Yan et al., 2013), and the semantic web (Berners-Lee et al., 2023).
To investigate data, researchers use specific data mining approaches, such as classification, regression, clustering, and anomaly detection (Basu & Meckersheimer, 2007; Carreno et al., 2020; Lin & Cohen, 2010). Time series anomaly detection has also garnered attention in recent years, due to its growing importance in fields such as urban planning, cybersecurity, healthcare risk analysis, and engine development (Zamanzadeh Darban et al., 2024). It combines, for instance, anomaly detection with clustering algorithms to identify acceleration events. Such analyses yield new semantic insights that can aid in the development of new engines, ensuring improved performance and reliability. One of the main goals in this area is compressing extensive datasets to extract essential information effectively (Tichomirov et al., 2023). By aggregating large datasets into manageable subsets, engineers can focus their attention on specific areas of engines, optimizing analysis time and resource allocation in engine development (Tichomirov et al., 2023).