The rapid advancements in sensing and measurement open the door for a wide range of signal-based applications across various fields. Moreover, the development of computing technologies and the rise of the Internet of Things (IoT) have paved the way to leverage machine learning (ML) within signal-based applications, offering new insights and achieving unprecedented levels of accuracy and efficiency. This merge between signal processing and ML is expected to play a major role in the next generations of sensor-enabled systems across various areas (D. K. Iakovidis et al., 2022). Thus, the integration of signal processing pipelines into ML models represents a critical intersection in developing the next generations of sensor-enabled systems, motivating the research community to address the role of signal processing in ML. This repository serves as a platform for posting a diverse collection of Python codes for signal processing that are relevant to signal-based ML applications. The codes facilitate various operations within a typical signal processing pipeline (pre-processing, processing, and application). The repository is associated with the paper entitled:

The paper, through its integrated-article approach, addresses the intersection of signal processing and ML by introducing the following contributions:

• Comprehensive Tutorial for a Broad Audience: The paper starts with a comprehensive tutorial on the fundamentals of signal processing that is aimed at readers from different domains, allowing the interested reader to develop a proper background before delving into the review.
• Application-Independent Approach: The paper adopts a broad, application-independent review, providing a comprehensive overview of signal processing in ML that is not shaped to a specific use case.
• End-to-End Overview of Signal Processing Workflow: The article thoroughly discusses the key tasks in a typical signal processing pipeline, grouping them under three main categories: preprocessing, processing, and application.
• Exhaustive Review with Novel Taxonomy: The paper conducts a detailed review of feature extraction techniques, presented and categorized through a new taxonomy that presents new insights and enriches the reader's understanding.
• Addressing Major Challenges and Potential Solutions: The paper identifies the primary challenges faced in implementing signal processing-based ML applications and discusses both existing and potential solutions.
• Bridging Theory with Practice: The paper addresses the practical application of signal processing in ML through two use cases. In the first use case, a spectral-based method is introduced for vibration-based condition monitoring of rolling bearings. In the second case, wavelet-energy analysis is utilized for epilepsy detection using EEG signals.
• Introducing this public repository of signal processing.

Each of these contributions has its own section in the paper. The sections are logically grouped into four parts. For reader convenience, the visual representation below shows various sections of the paper grouped under each part. Although these parts are organized to build knowledge progressively, they are written in a self-contained manner, allowing for selective reading based on interest or need.

For all inquiries or collaboration opportunities please contact:

Email : [email protected] or [email protected]
Github: SulAburakhia; Western OC2 Lab
Google Scholar: OC2 Lab; Sulaiman Aburakhia

If you find this repository useful in your research, please cite as:

Sulaiman Aburakhia, Abdallah Shami and George K. Karagiannidis, "On the Intersection of Signal Processing and Machine Learning: A Use Case-Driven Analysis Approach," arXiv:2403.17181, March 2024.

@ARTICLE{9855510,
  author={Aburakhia, Sulaiman A. and Shami, Abdallah and Karagiannidis, George K.},
  journal={ArXiv}, 
  title={On the Intersection of Signal Processing and Machine Learning: A Use Case-Driven Analysis Approach}, 
  year={2024},
  doi={arXiv:2403.17181}}