本地煮鸡:8080

# Introduction

# History

  • Ants, cavemen, and Early Recommender Systems
    • The emergence of critics
  • Information Retrieval and Filtering
  • Manual Collaborative Filtering
  • Automated Collaborative Filtering
  • The Commercial Era

# Information Retrieval and Filtering

# Information Retrieval

  • Static content base
    Invest time in indexing content

  • Dynamic information need
    Queries presented in "real time"

  • Common approach: TFIDF (Term Frequency-Inverse Document Frequency)
    Rank documents by term overlap
    Rank terms by frequency

  • Problem: The stream of information changes rapidly, but people only need the information related to their needs.
    (Information retrieval were reversed)

# Information Filtering

  • Reverse assumptions from IR
    Static information need
    Dynamic content base
  • Invest effort in modeling user need
    Hand-created "profile"
    Machine learned profile
    Feedback/updates
  • Pass new content through filters

# Manual Collaborative Filtering (CF)

  • Premise (前提)
    Information need more complex than keywords or topics: quality and taste
  • Small Community: Manual
    Tapestry: database of content and comments
    Active CF: easy mechanisms for forwarding content to relevant readers

# Recommendation Approaches

  • Non-Personalized and Stereotyped
    Popularity, Group Preference
  • Product Association
    People who like/bought X, also like Y
  • Content-Based
    Learn what I like (attributes)
  • Collaborative
    Learn what I like: use others' experience to recommend (multiple ways)

# Preferences and Ratings

  • Explicit (the preference got from asking users to give comments)
    Rating
    Review
    Vote
  • Implicit (the preference got by users' actions)
    Click
    Purchase
    Follow

# Predictions and Recommendations

  • Prediction
    • Estimates of how much user will like an item
      Often scaled to match some rating scale
      Often tied to search or browsing for specific products
    • Pro: helps quantify item
    • Con: provides something falsifiable (证伪的）
  • Recommendations
    • Suggestions for items user might like (or might fit what they are doing)
      Often presented in the form of "top-n" list
      Also sometimes just placed in front of the user
    • Pro: provides good choices as a default
    • Con: if perceived as top-n, can result in failure to explore (when the top few seem poor)
  • Explicitly(直白地) and Organically(自然逐步地)
    • Explicit recommend: recommend items directly
    • Organic recommend: recommend items gradually
    • Historical: Just for you (show the recommendation directly)
    • Today:
      Balance between explicit prediction(falsifiable) and coarser granularity(粗粒度)
      Balance between top-n and softer presentation(might be interesting)

# Taxonomy (分类学) of Recommenders

  • Dimensions of Analysis of Recommender System

    • Domain

      • Content to Commerce and Beyond
        News, information, "text"
        Products, vendors, bundles
        Matchmaking (other people)
        Sequences (e.g., music playlists)
      • One particularly interesting property
        New items (e.g., movies, books, ...)
        Re-recommend old ones (e.g., groceries, music)

    • Purpose

      • The recommendations themselves
        Sales
        Information
      • Education of user/customer
      • Build a community of users/customers around products or content

    • Recommendation Context

      • The user action when getting recommendation
      • The constrain approach of context to the recommender
        Groups
        Automatic consumption (vs. suggestion)
        Level of attention
        Level of interruption

    • Whose Opinions (Data)
      "Experts"
      Ordinary "phoaks"
      Normal people

    • Personalization Level

    • Privacy and Trustworthiness

    • Interfaces

    • Recommendation Algorithms

←