Recommendation and Personalization

NOTE: Article in Progress

personalized feeds, news

how much do we know about ourselves? Do what we think we know about ourselves is the same by what we do analysis on ourselves? How much of our reading (or watching) influence our thinking? An emerging field called personal informatics tries to answer these questions? But do we really need to know about ourselves? what about privacy? Is ignorance about self good?

past history

personal experiments? likes? bookmarks? Do they really tell anything about myself? Can this knowledge help me to improve myself?

user generated content

available information

why? information overload?

examples:

bookmarks

• title
• URL
• tag
• category
• keywords
• creation time
• modification time

commands

• input options and arguments
• time used

derivative commands

• use of pipes
• order of commands

tools and software applications

• commonly used menu options
• time used

feeds

• title
• URL
• tag
• category
• keywords
• creation time
• modification time
• application specific tags (important)

object tags and categories

• tags
• category (sub categories)
• creation time
• modification time

time gives history as well as evolution of information

justification/explanation for recommendation?

confirmation bias?

missing surprise element?

References

1. Personalization
2. Recommender System
3. Confirmation bias
4. Bias in algorithmic filtering and personalization, Engin Bozdag, Ethics and information technology, 2013, vol. 15, no 3, p. 209-227.