課程資訊
課程名稱
 實用決策科技
Practical Decision Technology 
開課學期
 107-2 
授課對象
 工學院  化學工程學研究所  
授課教師
 張良志 
課號
 ChemE5063 
課程識別碼
 524 U2240 
班次
  
學分
 3.0 
全/半年
 半年 
必/選修
 選修 
上課時間
 星期四6,7,8(13:20~16:20) 
上課地點
 普403 
備註
 化工選修課程。
限學士班三年級以上
總人數上限：30人 
Ceiba 課程網頁
 http://ceiba.ntu.edu.tw/1072ChemE5063_ 
課程簡介影片
 
核心能力關聯
 核心能力與課程規劃關聯圖
課程大綱
為確保您我的權利,請尊重智慧財產權及不得非法影印
課程概述

本課涵蓋數位決策所需的主要觀念及技術.課程分大部份:(1)數據科學及模式建構(Data Science. Machine Learning, Deep Learning), (2)模式優化及決策分析 (Decision Technology, Optimization)
課程重實際.主要單元都有對應的練習.所用軟件以Excel Data Analysis and Solver及Matlab Learner (Regression, Classification, Neural Network) 為主.
This course consists of two parts. The first part gives a broad coverage of modern modeling, Data Science, Machine Learning (ML) methods used for prediction and classification. It also includes Deep Learning Neural Networks for image and language processing. The second part provides a solid foundation of optimal decision making that can be applied to multiple business and engineering disciplines such as planning, profit maximization, business strategy setting, and logistics. Business and engineering applications will be discussed throughout the course.




 

課程目標
 學生可以學到並且直接應用工程及商業決策所需的實用技術.也可以見識到在化工業的大型應用為進一步研究或就業作準備.
除了硬性的技能之外本課也教學生(1)如何架構決策問題(2)如何把架構變成模式及(3)如何分析決策結果. 
課程要求
 大學數學, Excel, 基本化工程序知識. Matlab Learner (prior Matlab experience is not necessary) 
預期每週課後學習時數
  
Office Hours
  
指定閱讀
 1. "Management Science", 4th ed, S.G. Powell and K.R. Baker, Wiley, 2014.
Optional (Nice to Have):
2. "Hands-on Machine Learning with Scikit-Learn & TensorFlow", A. Geron, O'Reilly, 2018. 
參考書目
 1. “Deep Learning using Python”, F. Chollet, Manning, 2018
2. “Data Science for Business”, F Provost and T. Fawcett, O’Reilly, 2013
3. “Machine Learning with Matlab”, J. Smith, ?
4. “An Introduction to Statistical Learning”, 7th ed, G. James, D. Witten, T, Hastie, R. Tibshirani, Springer, 2017. (note: use R in examples)
5. “Predictive Analytics”, E. Siegel, Wiley, 2017
 
評量方式
(僅供參考)
    
課程進度
週次
日期
單元主題
第1週
 2/21  Course introduction and policy. State of art of Machine Learning.
Machine learning landscape and example.  
第2週
 2/28  Modeling framework. Introduction to optimization (Part I). Linear Regression and common modeling issues.  
第3週
 3/07  Regression and common modeling issues e.g. feature engineering, data partitioning, overfit and underfit issues etc. Large scale refinery planning application.  
第4週
 3/14  Classification methods-I and business applications (Logistic Regression),Principal Component Analysis (PCA) and its applications. 
第5週
 3/21  Classification methods (KNN, Logistic Regression, Decision Tree, Support Vector Machine) 
第6週
 3/28  Classification Performance measures and Clustering 
第7週
 4/09  Neural Network structures and calculation. Neural Net training.
 
第8週
 4/11  Deep Learning Neural Net (Convolutional NN and Recurrent NN for sequence data) 
第9週
 4/18  Mid-Term Exam. 
第10週
 4/25  Machine Learning Review for Midterm Exam
Mathematical foundation of optimization
 
第11週
 5/08  Linear Programming I: Translating business problem into LP formulation.  
第12週
 5/16  Linear Programming II Network Model, Applications in Model Predictive Control. 
第13週
 5/23  Non-Linear Programming in Engineering applications 
第14週
 5/23  Real-time Optimization in refinery and chemical processes.
Integer Programming  
第15週
 5/30  Integer Programming and Mixed Integer Programming 
第16週
 6/13  Uncertainty and Monte-Carlo Simulation. Review and Final Exam 
第17週
 6/13  Review before Final