1-1 About This Book ()

[chinese][english]

The goals of this books are:

• Cover the basic principles of data clustering (DC) and pattern recognition (PR).
• Exemplify the use of MATLAB for implementing DC and PR.
• Take real-world data as target applications for DC & PR.

The book is a practical guide for graduate students, researchers, as well as practitioner, with the following characteristics:

• Example-based tutorial: All the chapters include a number of examples, together with formal mathematical analysis and derivation.
• Emphasis on both theory and implementation: All algorithms covered in the text have accompanying MATLAB implementation, such that the uses can have hands-on experience to practice "learning by doing".
• Application oriented: Most of the examples take real-world data to verify the covered algorithms or methods. This enables the users to have a concrete idea of the gap between theory and implementation for real-world applications.

The depth of this book is designed for first-year graduate students. However, it is also suitable for upper-division of undergraduates if the lecturer put more emphasis on coding and implementation.

Speech and audio signal processing and recognition involves a fair amount of mathematics. We expect the readers to have taken the following prerequisites: Calculus, linear algebra, and probability.

This book was original written in Chinese. Therefore for some page, we have a link to the old Chinese version. However, it should be noted that the latest version is in English and there is no guarantee for the synchronization between English and Chinese versions.

If you want to cite this book, use the following format:

• Jyh-Shing Roger Jang, "Data Clustering and Pattern Recognition," available at the links for on-line courses at the author's homepage at http://mirlab.org/jang.

A set of loosely related slides is available here.

本書內容的重點如下：

• 介紹資料群聚（Data Clustering，簡稱 DC）與樣式辨認（Pattern Recognition，簡稱 PR）的基本原理。
• 說明如何以 MATLAB 進行 DC 和 PR 的程式碼實作。
• 以實際生活中的資料來說明 DC & PR 的各種相關應用。

筆者本著寫書著作的一貫宗旨，期望本書能達到下列終極目標：

• 範例式的教學：以簡單的範例來說明基本概念，然後再輔以正式的數學分析與推導。
• 理論與實作並重：所有的演算法都附有 MATLAB 的程式碼，讓使用者能夠穩紮穩打、Learning by Doing。
• 應用導向：所有的範例、理論與程式碼，最後都會用在現實世界中的應用，以讓讀者親自感受到各種演算法的長處和短處，以及程式碼實作方面可能遇到的困難。

當然，本書無法包山包海，所介紹的方法，都是比較偏向 DC & PR 的基本面，對一般碩士班研究生而言，應該已經夠用，但對於博士班研究生而言，可能尚嫌不足。不過無論讀者的背景為何，本書都可以做為 DC & PR 的入門書籍，並適合用於大學和研究所的相關課程。

當然，本書還是需要一些數學背景，才比較容易吸收，因此若做為課程教科書，修課同學應該先修過下列課程：基本微積分、線性代數、機率。

如果你在研究論文內要引用此書，可採用下列寫法：

• 張智星，"資料群聚與樣式辨認"，網路線上課程，可由作者之網頁 http://mirlab.org/jang連結到此線上課程。

---

Data Clustering and Pattern Recognition (資料分群與樣式辨認)