授業の形態
|
講義
|
|
アクティブラーニング
|
学生が議論する、学生が自身の考えを発表する、学生が文献や資料を調べる
|
|
授業内容と方法
|
*1: This class will open for below 4 classes: old curriculum: 「データマイニング論」 and "Data Mining Theory", new curriculum: 「データマイニング特論」 and "Advanced Data Mining", *2: This class will include Japanese students. *3: The class will be conducted remotely, and the Zoom URL for the first session will be provided in Webclass.
This class consists of 2 parts. We learn about a general process to be ready for re-useable knowledges by data mining theories and about machine learning models as a mining method. Everyday has about four persons have to explain/introduce about assigned references, basic theory from part 1 and applications (your research) from part 2.
|
|
URGCC学習教育目標
|
専門性
|
|
達成目標
|
+ You can explain about general process of data mining and some models in machine learnings. [Part 1] (Expertise) 3. Applied Subjects + You can explain/introduce about newly conference papers. [Part 2] (Expertise) 3. Applied Subjects + You can point out and discuss matters of concern from a fair perspective while being interested in each model and case and thinking independently. [Part 1, 2] (Expertise, Creativity, Ethics) 5. Related Subjects
Note: + The numbers in parentheses at the end of each item indicate the correspondence with URGCC-advanced (http://www.ged.skr.u-ryukyu.ac.jp/divisions/division-2/kaihatusitu). + The numbers also show the correspondence with the curriculum policy (https://www.u-ryukyu.ac.jp/admissions/3policy/gra_curriculumpolicy/#cat13).
|
|
評価基準と評価方法
|
You must explain/introduce/discuss about assigned references. presentation (60%), Q&A and discussions (20%), final report (20%).
The presentation will be adjusted according to the number of participants, but it will be allocated about three or four times for Part 1, one time for Part 2. Q&A refers to "questions to other presenters". Please note that even if you get a perfect score in the presentation and presentation documents, it will be only 80%. Voluntary participation in discussions during the weekly Q & A session is required.
|
|
履修条件
|
Math (especially Linear Algebra and Statistics), experiences of programming.
|
|
授業計画
|
Week 1 : Guidance
====================================== Part 1: Introduction to data mining and machine learning Week 2 ・Chap. 1, What's it all about? ・Chap. 2, Input: concepts, instances, attributes Week 3 ・Chap. 3, Output: knowledge representation ・Chap. 4, Algorithms: the basic methods, sec. 4.1 to 4.5 Week 4 ・cont., Chap. 4, Algorithms, section 4.6 to remains ・Chap. 5, Credibility: evaluating what's been learned Week 5 ・Chap. 6, Trees and rules ・Chap. 7, Extending instance-based and linear models Week 6 ・Chap. 8, Data transformations, section 8.1 to 8.3 ・Cont., section 8.4 to remains Week 7 ・Chap. 9, Probabilistic methods, section 9.1 to 9.5 ・Cont., section 9.5 to remains Week 8 ・Chap .10, Deep learning, section 10.1 to 10.3 ・Cont., section 10.4 to remains Week 9 ・Chap. 11, Beyond supervised and unsupervised learning ・Chap. 12, Ensemble learning Week10 ・Chap. 13, Moving on: applications and beyond
====================================== Part 2: Discussion about applications (your research) Week 11 ~ 15 ・two or three persons per week. ・if we have extra time, let's read conference paper.
|
|
事前学習
|
read carefully references.
|
|
事後学習
|
read carefully references.
|
|
教科書にかかわる情報
|
|
978-0128042915
|
we will read the part 1 of this book to learn the data mining theory.
|
Ian H. Witten ... [et al.]
|
Morgan Kaufmann
|
2017
|
|
|
|
|
教科書全体備考
|
|
|
参考書にかかわる情報
|
|
9781787125933
|
we will read the part 1 of this book to learn the data mining theory.
|
Sebastian Raschka, Vahid Mirjalili
|
Packt Pub.
|
2017
|
|
|
|
|
参考書全体備考
|
|
|
使用言語
|
英語
|
|
メッセージ
|
* In part 2, if we have extra time, we will read some best papers on below conferences in recent 5 years. But I hope that you suggest us any related papers in your interest.
+ mainly application examples. IEEE/WIC/ACM International Conference on Web Intelligence (IEEE/WIC/ACM WI) ACM International Conference on Web Search and Data Mining (ACM WSDM) ACM Special Interest Group on Information Retrieval (SIG-IR)
+ mainly theoretical or technical papers. IEEE International Conference on Data Mining (ICDM) ACM SIGKDD Conference on Knowledge Discovery and Data Mining (ACM KDD)
|
|
オフィスアワー
|
room: 705, eng. bldg. #1 office hour: 12:50-14:20 on Friday (in progress)
|
|
メールアドレス
|
この項目は教務情報システムにログイン後、表示されます。
|
|
URL
|
http://ie.u-ryukyu.ac.jp/~tnal/2023/adm/
|
|
|