Type of Credit: Elective
Credit(s)
Number of Students
本課程與國泰人壽, 新光人壽產學合作, 由公司的業師說明金融業實際需要解決的問題, 資料集, 再由課程中學習到的方法來解決並在期末報告展現成果. 修課同學須已經具備相當程度的 python /Matlab 程式撰寫能力.
能力項目說明
This course uses hands-on approach and utilize commercial software (MATLAB) and open source platform (GitHub). The students should have great interests in applications in Fintech and have solid background in mathematics (linear algebra, calculus, probability theory and optimization) and programming, especially in Python.
教學週次Course Week | 彈性補充教學週次Flexible Supplemental Instruction Week | 彈性補充教學類別Flexible Supplemental Instruction Type |
---|---|---|
課程進度:
9/13 Introduction of the course
9/20 Introduction of Machine Learning and Artificial Intelligence/The Machine Learning Landscape, chap 1 of Hands-on ML
9/27 Classification, chap 3 of Hands-on ML
10/4 業師說明要解決的問題以及資料集 I (國泰人壽)
10/11 業師說明要解決的問題以及資料集 II (國泰人壽)
10/18 MATLAB基本操作: 說明程式開發的基礎,變數宣告、函數使用、變數格式、流程控制、繪圖等。
10/25 業師說明要解決的問題以及資料集 III (新光人壽)
11/1 Support Vector Machines (SVMs), Decision Trees
11/8 Machine Learning cases in MATLAB: 信用評分卡建構: 課程說明使用MATLAB建構信用評分卡的流程,包含自動分箱(Binning),視覺劃分箱結果觀察分箱好壞並調整分箱,最後建構邏吉斯回歸以預測違約機率並建構評分卡。
11/15 Final project proposal
11/22 Ensemble Learning and Random Forests, chap 7 of Hands-on ML
11/29 Unsupervised learning
12/6 Deep learning
12/13 Deep learning深度學習預測信用違約 : 示範如何建立、訓練、比較不同深度學習模型以預測信用違約機率,運用MATLAB中的Deep Network Designer來建立並訓練深度學習模型12/21 Final project presentations I
12/28 Final project presentations I
1/4 Final project presentations II
Final Project Proposal 25%
Final Project 75%
書名 Book Title | 作者 Author | 出版年 Publish Year | 出版者 Publisher | ISBN | 館藏來源* | 備註 Note |
---|