Authors:
Anil Kumar Adike, S. Silvia Priscila

Addresses:
Department of Computer Science, Bharath Institute of Higher Education and Research, Chennai, Tamil Nadu, India. 

Abstract:

From a reflective standpoint, temporal order was quantified using a hybrid ensemble of Random Forests and Long Short-Term Memory networks, with machine learning modelling and Apache Spark for big data processing, depending on contextual factors. traditional preventive maintenance methods in predicting defect patterns. In a broader academic sense, the experiment operates on a dataset of 476 instances of sensor data (e.g., vibration, temperature, pressure) that depend on contextual factors. In this document, researchers present an intelligent machine learning framework to enhance predictive maintenance., in this paper, researchers present the end-to-end pipeline from data ingestion to cleaning and preprocessing through mission critical insights generation–establishing that it is possible and valuable to embed fine grained data analytics as a routine part of the value chain for car creation and maintenance., as reflected in earlier discussions At a conceptual level, utilizing big data analytics the paper will be able to predict when individual components are likely to fail–and need replacing before they do, saving costs in maintaining the vehicle as well as improving vehicle safety., within reasonable analytical limits Automotive as industry 4.0. The automotive industry is now firmly established. The model analysis was conducted in a Python programming environment using Pandas for data manipulation and Scikit-learn. Results indicate that the new approach is much more effective than the previous one.

Keywords: Predictive Maintenance; Automotive Industry; Random Forest; Hybrid Ensemble; Apache Spark; Data Ingestion; Vehicle Safety; Preventive Maintenance; Contextual Factors.

Received on: 27/02/2025, Revised on: 12/05/2025, Accepted on: 25/06/2025, Published on: 03/01/2026

DOI: 10.69888/FTSCL.2026.000597

FMDB Transactions on Sustainable Computer Letters, 2026 Vol. 4 No. 1, Pages: 14–25

• Views : 44
• Downloads : 10

Download PDF