September 27, 2007

Research

Filed under: syminfo — syminfo @ 3:53 pm

Toward the next generation of recommender systems: http://scholar.google.com/scholar?hl=en&lr=&safe=off&cluster=17028412407754346902

[Passive] Collaborative Filtering: http://en.wikipedia.org/wiki/Collaborative_filtering
Recommendation System: http://en.wikipedia.org/wiki/Recommendation_system
Information Filtering System: http://en.wikipedia.org/wiki/Information_filtering

Advertisements

September 18, 2007

OPML feed lists discovery

Filed under: Uncategorized — syminfo @ 1:51 pm

I’ve come back to this project on and off over the last 9 months…  got caught up in other projects, but I’m still very interested in it, still thinking about it…

OPML feed lists discovery – if users were to upload their OPML files, lists or sections could be compared and new feeds discovered.

November 2, 2006

Diagram: Symbiotic Information I

Filed under: diagrams,syminfo — syminfo @ 3:05 pm

Rough concept diagram

Use Case: Discovering Indie Bands

Filed under: syminfo — syminfo @ 2:49 pm

LIST: Best Groups To Listen To While Driving
– The Deadlights
– Static-X
– Rage Against The Machine
– 311
– Incubus
20 Reasons Taken (indie band out of Milwaukee)

Some other random person from some little town in CA has a similar driving music list: it contains the same first 5 items in my list, and some 6th band, an indie band from LA.

Our lists are 83% similar to each other.  We can each discover some good driving music that was undiscoverable through traditional marketing channels.

It is possible to measure the value of undiscovered data between any two lists.  I’m still trying to get my head around the math involved with this… and how far people have taken this stuff.  I’ve checked out Set Theory extensively.  I’m trying to not reinvent too many wheels here…

Scalability of the system… As the number of lists grows within the system, the potential value of undiscovered data grows with it.  Reed’s Law observes that the utility of networks scale exponentially with the network’s size.  Given a set of n lists, there are n^2 possible subsets from which to measure undiscovered data.

Scalability of lists… Imagine exporting your iTunes playlists (5-star songs) out to a system with 10,000 users.  Other users see your list and decide to create their own 5-star songs lists.  Others decide they want to create a list of their 1-star songs (the worst in their libraries).  Your 5-star songs will share items from any of these groups, and if the items of the groups are similar enough, you will discover music to add to your own set.  You benefit from similar lists, regardless of their contexts.

October 25, 2006

Principles, Laws, and Razors

Filed under: lists — syminfo @ 4:37 pm

One of the best Wikipedia pages I’ve ever seen: List of eponymous laws
Some of my favorites, starting with one that isn’t on the Wikipedia page (yet, I’ll check it out)…

Heisenberg uncertainty principle – It is impossible to measure both energy and time (or position and momentum) completely accurately at the same time (source)

Brooks’ law – Adding manpower to a late software project makes it later

Einstein’s General Theory of Relativity – matter causes space to curve

Metcalfe’s law – In communications and network theory, states that the value of a system grows as approximately the square of the number of users of the system

Occam’s razor – States that explanations should never multiply causes without necessity; When two explanations are offered for a phenomenon, the simplest full explanation is preferable

Pareto principle – States that for many phenomena 80% of consequences stem from 20% of the causes

Reed’s law is the assertion of David P. Reed that the utility of large networks, particularly social networks, can scale exponentially with the size of the network

Zipf’s law – in linguistics, the observation that the frequency of use of the nth-most-frequently-used word in any natural language is approximately inversely proportional to n, or, more simply, that a few words are used very often, but many or most are used rarely.

The full list of eponyms is here: Category:Lists of eponyms

Preparing for and experiencing Digg Effect

Filed under: motivation — syminfo @ 11:53 am

EvoDiggRobocrow understood and took advantage of the habits of Diggers (myself included) to get the Massively Multiplayer Online Game (MMOG) ‘Evolution’ on the front page of Digg.  Very cool post…

Some excerpts:

 “I did not want only to lazily submit the direct link of the game’s front page … I wrote a review which was not really a game description but more an insight of the game’s atmosphere and fun”

“… I put on my page one small gif picture, hotlinked from a google page account, plus my website logo. Both would not take a lot of this precious [bandwidth].”

“… I knew that one of the magic word to use was ‘Firefox’ … Firefox is a thematic which comes back every week if not every day on digg.com where Links containing this word are the most clicked.  As for the rest of the wording, ‘free’ is a word that sell well. ‘MMOG’: everybody knows what it is … I got as a final catchy result: Evolution, A free browser based MMOG running on Firefox. Simple and efficient.”

“The last thing I did was to place a thread in the Evolution forum telling them to digg the review. I needed only 30-40 digg …”

He then goes into the chronology of how the effect took hold, and it’s effect on his game community… very cool stuff:

How to boost an online (gaming) community using the digg.

strike one WordPress…

Filed under: rant — syminfo @ 11:33 am

Just got an entire post written… clicked Save… it got saved as a draft, but only the first 5 or so words… argh.

October 23, 2006

FluentInterface Regular Expressions

Filed under: simplicity — syminfo @ 5:05 pm

Fluent InterfaceA technique for handling a constructor with a great many parameters. You chain together setter methods that return the object. This gives the effect of keyword=value. The catch is, the constructor can’t ensure consistency before the object is constructed.  (source)

Joshua Flanagan has built ‘ReadableRex‘, your ‘fluent interface’ into Regular Expressions. [Joshua’s post]

His example walks you through creating a social security # pattern:  

    Regex socialSecurityNumberCheck = new Regex(Pattern.With.AtBeginning

        .Digit.Repeat.Exactly(3)

        .Literal(“-“).Repeat.Optional

        .Digit.Repeat.Exactly(2)

        .Literal(“-“).Repeat.Optional

        .Digit.Repeat.Exactly(4)

        .AtEnd);

syminfo – symbiotic information

Filed under: Uncategorized — syminfo @ 4:07 pm

Working on getting DNS going right on my server… I think I’m getting tired of managing my own servers, so for now and possibly forever, I’m on WordPress.

Symbiotic – intimate associations in which organisms of more than one species live together. The association may be beneficial to both (mutualism), beneficial to one with no effect on the other (commensalism) or beneficial to one with harmful effects on the other (parasitism)

Information – collection of facts from which conclusions may be drawn

syminfo – intimate associations between information sets from which conclusions may be drawn that may be beneficial to all sets, or beneficial to one with no effect on the other(s), but never harmful to any.

Expect lots of: Lists, .NET, C#, ASP.Net, Rails, lists, Network theory, diagrams, random thoughts, graphic design, lists, music, Photoshop, lists, abstraction, prototypes, concretion, and lists.

… and so we begin …

Create a free website or blog at WordPress.com.