Below are some books that we have read, found very useful, and highly recommend. Let us know what you think of these books and if there are other books (related to machine learning and, preferably, connected to process industry) that we should include in this list.
Title: Fault Detection and Diagnosis in Industrial Systems (Amazon link)
About the book: Among the first books dedicated to fault detection and diagnosis techniques for modern chemical plants and focusses primarily on multivariate statistical process monitoring (MSPM).
The book is reasonably-sized (~250 pages) and when you are done with the book, you will have solid conceptual understanding of the classical MSPM techniques (PCA/PLS/ICA/FDA/CVA).
What we liked about it:
‘Easy’ to read
Easy to understand mathematical exposition
Clear explanation of conceptual underpinnings
Title: Principles of System Identification: Theory and Practice (Amazon link)
About the book: Perfect book for those new to system identification and looking to master this area. The book covers classical time series models (AR/MA/ARMA/ARIMA) and I/O models (FIR, ARX, ARMAX, ARIMAX, OE, BJ).
Running at around 800 pages, the book is bulky, but it will be worth your time!
What we liked about it:
Perfect combination of theory and practice; focus on conceptual understanding
Matlab implementation of every concept
Uses process industry-relevant examples for illustrations
Title: Multivariable System Identification For Process Control (Amazon link)
About the book: Designed for industrial practitioners, this book attempts to bridge the gap between theory and application of system identification methods. All aspects related to multivariate dynamic process identification are covered.
The book is reasonably-sized (~350 pages). If you work in process industry and are involved in dynamic process modeling, then you will find this book useful.
What we liked about it:
Coverage of techniques for experiment design for data collection
Practice-oriented coverage of popular PEM models (FIR, ARX, etc.) and Hammerstein/Wiener models
Presents industrial case-studies along with Matlab implementations
Title: Time Series Analysis: Forecasting and Control (Amazon link)
About the book: Provides a comprehensive coverage of the tools for modeling and analyzing time series and I/O systems. The book is well-suited for both academicians and practitioners.
Not much need to be said about this classic book. The authors are legends in the area of time series analysis. The book is bulky (~700 pages), but a worthy investment of your time.
What we liked about it:
Mathematical details are presented nicely
A to Z coverage of all concepts related to modeling of dynamic systems
Data used for illustrations are publicly available
Title: Industrial Approaches in Vibration-Based Condition Monitoring (Amazon link)
About the book: Vibration-based condition monitoring (VCM) is synonymous with predictive maintenance and this book can help you quickly gain working-level knowledge of this critical area.
The book is reasonably-sized (~250 pages) and when you are done with the book, you will have a good understanding of vibration data acquisition and signal processing. The book does not, however, cover ML-based VCM.
What we liked about it:
Beginner-level elucidation of VCM concepts
Written in tutorial style with focus on industrial implementations
Title: Intelligent Fault Diagnosis and Remaining Useful Life Prediction of Rotating Machinery (Amazon link)
About the book: Provides a comprehensive introduction of machine learning-based intelligent fault diagnosis and RUL prediction. Shows how to build advanced VCM solutions using ANNs, Deep Learning, SVMs, kNN, clustering, etc.
The book is reasonably-sized (~350 pages) and will help you get a good understanding of how to use advanced techniques for vibration signal processing and build a ML-based VCM solution.
What we liked about it:
Coverage of machine learning using vibration signals
Coverage of advanced techniques for vibration signal processing
Showcases applications of covered techniques on real systems
Title: Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow (Amazon link)
About the book: An introductory book on modern machine learning providing comprehensive coverage of techniques ranging from linear regression to deep learning. Provides step-by step guidance to build models for computer vision, natural language processing, etc.
This is one of the most popular book for introductory machine learning in Python. The book does not go too much into the mathematical details but provides adequate insights into each technique. It is a good book to build your ML foundations.
What we liked about it:
Suitable for both beginners and expert data scientists
Uses popular Python frameworks (Scikit-Learn, Keras, and TensorFlow) for code implementations
Easy-to-read. Focus on conceptual understanding
Title: Python Machine Learning: Machine Learning and Deep Learning with Python (Amazon link)
About the book: A step-by-step tutorial to machine learning and deep learning with Python. It acts as both a step-by-step tutorial. Packed with clear explanations, visualizations, and working examples.
This is another very popular book for introductory machine learning in Python. The book goes into the details of each technique and shows you how to implement them from scratch! It is a perfect book to understand the nitty-gritty of algorithms behind popular ML techniques.
What we liked about it:
Clear and intuitive explanations
Detailed code implementations
Coverage of all modern ML techniques
0 Comments