查詢教學大綱與進度

教學大綱與進度

課程基本資料：

學年期課號課程名稱階段學分時數修教師班級人撤備註

114-1 352488 人工智慧跨域商情預測之應用 1 3.0 3 ★ 陳擎文經管三 40 0 　

教學大綱與進度：

教師姓名陳擎文　教師諮商時間(Office Hours)

Email ccw@ntut.edu.tw

最後更新時間 2025-06-14 17:31:21

課程大綱本課程將介紹人工智慧在各種商情預測的技應用，做好商情預測有利於確保企業發展方向的正確性，也能讓企業更能抓住市場脈動而滿足消費者的需求不會造成太多的庫存堆積，保持市場供需的動態平衡。本課程規劃所欲培育人才所需之教學目標包括：（1）學生有能力使用python寫各種商業智能程式，包括使用pandas模組裡進行商業數據分析與視覺化。使用numpy，sklearn機器學習模組來處理迴歸分析問題。使用keras模組來進行深度學習的商業預測應用。（2）實作多種商情預測範例，包括：電商數據分析、市場佔有率分析、市場趨勢分析、細分市場分析、消費者需求分析、商品推廣方案分析、競品分析、銷售預測、庫存預測。 This course will introduce the technological application of artificial intelligence in various business forecasts to help students understand the current status of various applications of AI in business forecasts. The teaching objectives required for cultivating talents in this course planning include: (1) Students are able to use python to write various business intelligence programs, including using pandas modules for business data analysis and visualization. Use numpy, sklearn machine learning module to handle regression analysis problems. Use keras modules for business forecasting applications of deep learning. (3) Implement a variety of business forecast topics.

課程進度第1週：人工智慧跨域商情預測各種應用的內容與現況，包括：市場需求預測、市場供應預測、市場供求動態預測、企業市場佔有率預測、以及市場價格預測等。第2週：人工智慧第1步：資料分析之數據截取與查詢（pandas-1）第3週：數據處理基礎練習，及數據的3種格式：DataFrame，array矩陣，list串列數組第4週：人工智慧的基礎與數學理論第5週：人工智慧第2步：機器學習：使用numpy模組來學迴歸分析 ptyhon人工智慧『簡易3步驟』程式寫法(簡易基礎13題) 第6週：機器學習：使用sklearn模組來建立線性/非線性迴歸模型 ptyhon人工智慧『入門6步驟』程式寫法(入門13題，加上:評估model好壞，驗證overfitting，關聯度) 第7週：機器學習：使用sklearn模組來建立多元線性迴歸模型資料集的視覺化(3種方法) 第8週：期中考第9週：商業分類問題實戰分類問題的機器學習：建立『邏輯斯模型(logistic regression model)，KNN模型』來預測二元分類問題第10週：商業二元分類問題建立『決策樹，隨機森林，支持向量機算法』來預測二元分類問題第11週：非監督式學習應用集群分析/分群分析，建立『K-means分群』模型來把『自動分群』降維簡化分析(dimension reductoin)，建立『主成份分析PCA』降維模型，來把特徵參數簡化降維第12週：高階商業分類模型與視覺化第13週：深度學習基礎（Keras框架）第14週：深度學習分類實戰人工智慧第3步：深度學習：使用keras模組來進行人工智慧預測第15週：數據預處理自動化資料預處理，標準化(Standardization), 特徵縮放scaler 第16週：特徵選取策略 8種特徵選取feature selection方法用管道器封裝中間的全部流程：(x-->【管道器pipeline】-->y) 第17週：專案實作：消費者需求分析，用戶流失預測，銷售預測、庫存預測第18週：期末考這門課全程寫程式，請同學們上課要繫帶筆記本電腦 ----------------- Weekly Schedule Week 1: Cross-Domain AI Applications in Business Forecasting Current practices: Market demand/supply/dynamic equilibrium forecasting, market share prediction, and price trend analysis. Week 2: AI Step 1 – Data Acquisition & Query (Pandas-1) Data extraction from CSV/Excel, conditional filtering (df.query()). Week 3: Data Processing Fundamentals Three data structures: DataFrame (tabular), Array (matrix), and List (sequential). Week 4: Mathematical Foundations of AI Key concepts: Loss functions, gradient descent, and feature engineering. Week 5: AI Step 2 – Machine Learning with NumPy Simplified 3-step regression modeling (13 basic exercises). Week 6: Linear/Nonlinear Regression with Scikit-learn Enhanced 6-step workflow: Model evaluation, overfitting validation, and feature correlation. Week 7: Multivariate Regression & Data Visualization Three visualization methods: Pandas built-in, Matplotlib, and Seaborn. Week 8: Midterm Exam Week 9: Classification Models for Business Problems Binary classification: Logistic Regression and K-Nearest Neighbors (KNN). Week 10: Advanced Binary Classification Decision Trees, Random Forest, and SVM for customer value prediction. Week 11: Unsupervised Learning Applications Clustering: K-means for automated segmentation. Dimensionality reduction: PCA for feature simplification. Week 12: Advanced Classification & Visualization Model interpretation with SHAP values and confusion matrices. Week 13: Deep Learning Basics (Keras Framework) Neural network architecture design. Week 14: Deep Learning for Classification Case studies: Diabetes prediction and Iris species classification (One-Hot Encoding). Week 15: Automated Data Preprocessing Standardization, feature scaling (MinMaxScaler), and pipeline construction. Week 16: Feature Selection Strategies 8 methods: Filter/Wrapper/Embedded approaches. End-to-end pipeline encapsulation: sklearn.pipeline. Week 17: Capstone Project Options: Consumer demand analysis, churn prediction, sales/inventory forecasting. Week 18: Final Exam

評量方式與標準 1.上課分數：20% 2.作業分數：30% 3.期中考：25% 4.期末考：25% 這門課全程寫程式，請同學們上課要繫帶筆記本電腦 --- 1.The prerequisite for enrolling in this course is the completion of the Department of Business Management's course titled **"Interdisciplinary Applications of Artificial Intelligence in Business Management."** 2.This course exceeds my capacity if it has more than 45 students; I will no longer accept students if it goes beyond that. 3.socreing method: Class Participation: 20% Assignments: 30% Midterm Exam: 25% Final Exam: 25% This course involves programming throughout; please bring a notebook computer to class. 3.This course is a selective course and can only be taken if you have previously completed the required course [Applications of Artificial Intelligence in Cross-Domain Business Management]; otherwise, you cannot enroll.

使用教材、參考書目或其他 【遵守智慧財產權觀念，請使用正版教科書，不得使用非法影印教科書】

使用外文原文書：否
1.自編教學網站：（1）https://acupun.site/lecture/predict/ （2）https://acupun.site/lecture/pandas 2.建議書籍：商用大數據分析，梁直青，全華圖書一行指令學Python：用機器學習掌握人工智慧，徐聖訓，全華圖書 3.參考書目：（1）Python商業數據分析：零售和電子商務案例詳解，作者：零一，電子工業出版社（2）一行指令學Python－用Pandas掌握商務大數據分析，作者：徐聖訓，出版社：全華圖書

課程諮詢管道 1.可以在office hours時段來諮詢(科研大樓537室)
3.老師的email：ccw@ntut.edu.tw

課程對應SDGs指標 SDG4：優質教育（Quality Education）
SDG9：產業創新與基礎設施（Industry, Innovation and Infrastructure）
SDG11：永續城市與社區（Sustainable Cities and Communities）

課程是否導入AI ● 使用生成式AI作為教學過程的輔助工具（Use generative AI tools as an aid to the teaching process）
● 鼓勵學生使用生成式AI工具（Encourage students to use generative AI tools）
● 使用生成式AI進行課堂即時互動（Use generative AI for real-time classroom interaction）
● 簡單介紹AI的基本概念（Briefly introduce the basic concept of AI）
● 深入探討AI的原理與應用（Explore the principles and applications of AI in depth）

備註
因應疫情發展，本學期教學及授課方式請依照學校網頁所公布之訊息為準：
(https://oaa.ntut.edu.tw/p/404-1008-98622.php?Lang=zh-tw)

同學們上課要繫帶筆記本電腦

因應疫情發展，本學期教學及授課方式請依照學校網頁所公布之訊息為準：
(https://oaa.ntut.edu.tw/p/404-1008-98622.php?Lang=zh-tw)
若因疫情而暫停實體上課時，將使用學校的Microsoft Teams 課程名稱