Notes on codes, projects and everything
ann
-
Approximate Neighbour Search in Multiple Dimensions
In the previous post, I re-implemented Annoy in 2D with some linear algebra maths. Then I spent some time going through some tutorial on vectors, and expanded the script to handle data in 3D and more. So instead of finding gradient, the perpendicular line in the middle of two points, I construct a plane, and find the distance between it and points to construct the tree.
-
Re-implementing Approximate Nearest Neighbour Search
Recently I switched my search code to Annoy because the input dataset is huge (7.5mil records with 20k dictionary count). It wasn’t without issues though, however I would probably talk about it next time. In order to figure out what each parameters meant, I spent some time watching through the talk given by the author @fulhack.
Random Posts
-
Resource Definition Framework
As the name implies, Resource Definition Framework, or RDF in short, is a language to represent information about resources in world wide web. Information that can be represented is mostly metadata like title (assuming the resource is a web-page), author, last modified date etc. Besides representing resource that is network-accessible, it can be used to represent things that cannot be accessed through the network, as long as it can be identified using a URI.
-
Regression with 2-means clustering Annoy (non-scientific)
So apparently Annoy is now splitting points by using the centroids of 2 means clustering. It is claimed that it provides better results for ANN search, however, how does this impact regression? Purely out of curiosity, I plugged a new point splitting function and generated a new set of points.
(more…) -
Linked Data Principle
Semantic Web is not just about putting data on the web, but also making links to allow a person as well as a machine to explore the web of data. Links are made in the web of data connects arbitrary things together as described by RDF as opposed to links in the web of hypertext, where links connects to only web-resources. Linkage of arbitrary things then allow related things to be found while performing search.
-
Over-doing a task: reading CSV file in #golang
I was invited to try Go (the programming language, not that board game) a few months ago, however I didn’t complete back then. The main reason was because it felt raw, compared to other languages that I know a fair bit better (for example Ruby). There was no much syntatic sugar around, and getting some work done with it feels “dirty”.
-
Extending Zend Framework
I like how Kohana 3 organizes the classes, and I thought the same thing may be applied to my Zend Framework experimental project. Basically what this means is that I can name the controller class according to PEAR naming convention, and deduce the location of the file by just parsing the class name.
