UMN CSCI 5123 Recommender System

• Syllabus for CSCI 5123 (001) Recommender Systems (Fall 2019)
  • Assignments: CSCI 5123 (001) Recommender Systems (Fall 2019)
• Coursera
  • Recommender Systems | Coursera (Specialization)
    1. Introduction to Recommender Systems: Non-Personalized and Content-Based | Coursera
    2. Nearest Neighbor Collaborative Filtering | Coursera
    3. Recommender Systems: Evaluation and Metrics | Coursera
    4. Matrix Factorization and Advanced Techniques | Coursera
    5. Recommender Systems Capstone | Coursera

The original class use Java as primary language, but I will try to do it in Python.

• lenskit/lenskit-hello: Hello World for LensKit. (deprecated)
• lenskit/lkpy: Python recommendation toolkit
  • LensKit — LensKit 0.13.1 documentation
  • Paper: LensKit for Python

Schedule

Coursera Recommender System Specializations

Include some reference of the Main Course

1. Introduction to Recommender Systems: Non-Personalized and Content-Based

• Preparation for Today's Class (9/9) - Basically the Week 1 content
• Preparation for Today's Class (9/16) - Basically the Week 2 content

• Week 1
  • Preface
  • [X] Intro to Recommender Systems (1)
  • [X] Intro to Course and Specialization (2)
  • [X] Notes on Course Design and Relationship to Prior Courses
  • Introducing Recommender Systems
  • [X] Movielens Tour (3)
  • [X] Preferences and Ratings (4)
  • [X] Predictions and Recommendations (5)
  • [X] Taxonomy of Recommenders I (6)
  • [X] Taxonomy of Recommenders II (7)
  • [X] Tour of Amazon.com (8)
  • [X] Recommender Systems: Past, Present and Future (9)
  • [X] Introducing the Honors Track (10)
  • [X] Honors: Setting up the development environment (11)
  • [X] About the Honors Track
  • [X] Downloads and Resources
  • My Alternative
  • Getting Started — LensKit 0.13.1 documentation
  • Getting Started with LensKit for Python - Colaboratory
  • Local notebook
• Week 2
  • Non-Personalized and Stereotype-Based Recommenders
  • [X] Non-Personalized and Stereotype-Based Recommenders (1)
  • [X] Summary Statistics I (2)
  • [X] Summary Statistics II (3)
  • [X] Demographics and Related Approaches (4)
  • [X] Product Association Recommenders (5)
  • [ ] Assignment #1 Intro Video
  • [ ] Assignment Intro: Programming Non-Personalized Recommenders
  • [X] Additional Reading of the Course Preparation (Notes)
  • [X] UX Research: 5 Requirements for the ‘Ratings Distribution Summary’ on the Product Page (65% Get it Wrong) – Articles – Baymard Institute (backup)
  • [X] How Not To Sort By Average Rating – Evan Miller (backup)
  • [X] How Hacker News ranking algorithm works | by Amir Salihefendic | Hacking and Gonzo | Medium (backup)
  • [X] How Reddit ranking algorithms work | by Amir Salihefendic | Hacking and Gonzo | Medium (backup)
  • Honors Programming Assignment: Programming Non-Personalized Recommenders
  • [ ] Non-Personalized Recommender Assignment
• Week 3
  • Content-Based Filtering Using TFIDF
  • [X] Introduction to Content-Based Recommenders (1)
  • [ ] TFIDF and Content Filtering
  • [ ] Content-Based Filtering: Deeper Dive
  • Advanced Content-Based Techniques and Interfaces
  • [ ] Entree Style Recommenders -- Robin Burke Interview
  • [ ] Case-Based Reasoning -- Interview with Barry Smyth
  • [ ] Dialog-Based Recommenders -- Interview with Pearl Pu
  • [ ] Search, Recommendation, and Target Audiences -- Interview with Sole Pera
  • [ ] Beyond TFIDF -- Interview with Pasquale Lops

Assignments

Homework #1

• Homework #1 -- Non-Personalized and Content-Based Recommenders

Resources

• LaTeX Math Symbols

Code

• LensKit — Welcome
  • lenskit/lkpy: Python recommendation toolkit
  • LensKit Recommender Toolkit
• shenweichen/DeepCTR: Easy-to-use,Modular and Extendible package of deep-learning based CTR models .
• cheungdaven/DeepRec: An Open-source Toolkit for Deep Learning based Recommendation with Tensorflow.
• lyst/lightfm: A Python implementation of LightFM, a hybrid recommendation algorithm.
• rixwew/pytorch-fm: Factorization Machine models in PyTorch
• ZiyaoGeng/Recommender-System-with-TF2.0: Recurrence the recommender paper with Tensorflow2.0
• google-research/recsim: A Configurable Recommender Systems Simulation Platform

Blog

• Twitter Recommendation Algorithm
  • A new era of transparency for Twitter
  • Twitter's Recommendation Algorithm
  • twitter/the-algorithm: Source code for Twitter's Recommendation Algorithm
  • twitter/the-algorithm-ml: Source code for Twitter's Recommendation Algorithm
  • Twitter GitHub

Notes

• ShaoQiBNU/videoRecTips: 短视频推荐优化经验总结

Book

• 深度學習推薦系統 - 王喆

Trouble Shooting

Windows filename too long

1. regedit > HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem > LongPathsEnabled (set to 1)
2. gpedit.msc (require Windows Pro) > Computer Configuration > Administrative Templates > System > Filesystem > Enable Win32 long paths (set to enabled)