近年來物聯網的發展，印刷電路板的市場也隨之增加。因此印刷電路板製造商需要更有效率的方式來加速製造設計流程。電子零件圖樣分類能夠幫助分析製造規則的分析以及完成一部份電子零件圖樣的設計自動化。印刷電路板上的電子零件圖樣在印刷電路板中為一個重要的角色，電子零件圖樣中存有大量電子零件的資訊，其中包含零件的外型、製造時的公差、印刷電路板製造廠的設計規則等等。不同的電路板製造廠面對相同的零件會根據不同的設計規則設計出不同的電子零件圖樣。因此，藉由將零件圖樣分類可以幫助製造商分析相同封裝類型的零件圖樣設計規則。在這篇論文中，我們結合電子零件圖樣以及圖樣的檔案名稱對電子零件圖樣進行分類，能得到較高的精準度以及完成一部份的設計自動化。

The market of Printed Circuit Boards (PCB) is growing fast with the popula-tion of the Internet of Things. Therefore, PCB manufacturers require an effective footprint design method to accelerate the PCB manufacturing processes. Foot-print classification that helps PCB footprint design rules analyzing and footprint design can achieve a part of footprint design automation. PCB electric component footprint is one of the important parts in PCB development, which contains lots of information, such as component’s outline, tolerance, and PCB manufacturer’s design behaviors. Since different manufacturers follow different design rules to produce components, the corresponded footprint for same components might be different. Thus, by classifying the footprint, we can analyze the design rules for those footprints in same package. PCB footprint classification becomes necessary. In this paper, we combine both PCB footprint classification and PCB footprint file name classification to classify PCB footprints. Through the proposed method, we can classify the footprint libraries with higher accuracy so as to achieve a part of footprint design automation.

Abstract i
Abstract ii
Acknowledgment iii
1 Introduction 1
2 Background and Problem Formulation 4
2.1 Process of Footprint Designing . . . . . . . . . . . . . . . . . . . 4
2.2 Difficulties in PCB Footprint Classification . . . . . . . . . . . . . 5
3 Proposed Methodology 7
3.1 Overall Architecture . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 PCB Footprint Classification . . . . . . . . . . . . . . . . . . . . 7
3.3 File Name Classification . . . . . . . . . . . . . . . . . . . . . . . 9
3.4 Combines Two Classification Results . . . . . . . . . . . . . . . . 9
4 Experimental Evaluation 11
iv
4.1 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . 12
5 Conclusions 15

[1] L. McSherry, “Image showing populated and unpopulated TSOP land patterns,”
https://zh.wikipedia.org/wiki/Footprint#/media/File:TSOP_
LandPatterns_Pop+Unpop.JPG.
[2] Te Connectivity, https://www.te.com/usa-en/product-1-188275-6.html.
[3] Y.-J. Wang, Y.-T. Chen, Y.-S. F. Jiang, M.-F. Horng, C.-S. Shieh, H.-Y. Wang,
J.-H. Ho, and Y.-M. Cheng, “An artificial neural network to support package
classification for smt components,” in 2018 3rd International Conference on
Computer and Communication Systems (ICCCS), pp. 130–134, IEEE, 2018.
[4] T. Chen, T. He, M. Benesty, V. Khotilovich, and Y. Tang, “Xgboost: extreme
gradient boosting,” R package version 0.4-2, pp. 1–4, 2015.
[5] A. Aizawa, “An information-theoretic perspective of tf–idf measures,” Information
Processing & Management, vol. 39, no. 1, pp. 45–65, 2003.
[6] K. Mitzner, Complete PCB design using OrCAD Capture and PCB editor.
Newnes, 2009.
[7] M. W. Al Nabki, E. Fidalgo, E. Alegre, and R. Alaíz-Rodríguez, “File name
classification approach to identify child sexual abuse.,” in ICPRAM, pp. 228–
234, 2020.
[8] Y. Kim, “Convolutional neural networks for sentence classification,” arXiv
preprint arXiv:1408.5882, 2014.
[9] scikit-learn: machine learning in Python —scikit-learn 0.23.2 documentation
https://scikit learn.org/stable/.