Title: Time series models for web service activity prediction

Authors: Kambhampati Mukta; Sandhya Harikumar

Addresses: Department of Computer Science and Engineering, Amrita School of Computing, Amrita Vishwa Vidyapeetham, Amritapuri – 690525, India ' Department of Computer Science and Engineering, Amrita School of Computing, Amrita Vishwa Vidyapeetham, Amritapuri – 690525, India

Abstract: Web service providers have to be very vigilant in offering their services to their clients and ensure that there are no glitches. In this research, we propose a method for anticipating the load of user activities at various intervals of time that can help in identifying the best time for maintaining the various software upgrades. Machine learning approach for predicting the user traffic is leveraged based on various time series analysis techniques and long short-term memory (LSTM). Further, we provide a graphical visualisation of user traffic at regular intervals and notify the stakeholders if the traffic is more than the threshold. The contribution of the work lies in embedding time series analysis with good visualisation for real-time monitoring, querying the traffic condition, and predictive analysis. The results are validated using the evaluation metric mean absolute percentage error (MAPE) and the visualisations are rendered using Grafana visualisation tool.

Keywords: machine learning; time series analysis; ARIMA; SARIMA; Grafana; Prometheus; long short-term memory; LSTM.

DOI: 10.1504/IJCSE.2024.139692

International Journal of Computational Science and Engineering, 2024 Vol.27 No.4, pp.406 - 413

Received: 17 Sep 2022
Accepted: 24 Jun 2023
Published online: 05 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article