Notes on codes, projects and everything
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…)After a year and half, a lot of things changed, and annoy also changed the splitting strategy too. However, I always wanted to do a proper follow up to the original post, where I compared boosting to Annoy. I still remember the reason I started that (flawed) experiment was because I found boosting easy.
(more…)While following through the Statistical Learning course, I came across this part on doing regression with boosting. Then reading through the material, and going through it makes me wonder, the same method may be adapted to Erik Bernhardsson‘s annoy algorithm.
(more…)Traversing a tree structure often involves writing a recursive function. However, Python isn’t the best language for this purpose. Therefore I started flattening the tree into a key-value dictonary structure. Logically it is still a tree, but it is physically stored as a dictionary. Therefore it is now easier to write a simple loop to traverse it.
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.
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.
So I first heard about Panda probably a year ago when I was in my previous job. It looked nice, but I didn’t really get the chance to use it. So practically it is a library that makes data looks like a mix of relational database table and excel sheet. It is easy to do query with it, and provides a way to process it fast if you know how to do it properly (no, I don’t, so I cheated).
So my cheat with dask worked fine and dandy, until I started inspecting the output (which was to be used as an input for another script). While the script seemed to work fine, however when I started to parse each line I was hit with some funny syntax errors. After some quick inspection I found some of the lines was not printed completely.
Long long time ago when I was working with Prolog, I was introduced to list. Unlike arrays in most popular programming languages, we weren’t really able to access to a particular member directly. Every list is constructed in a chain-like structure.
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.
After publishing the previous note on setting up my development environment, I find myself spending more time in the CLI (usually via SSH from host). Then I find myself not needing all the GUI apps in a standard Ubuntu desktop environment so I went ahead and set up a new environment based on Ubuntu Quantal server edition beta-1. For some reason my network stopped working and didn’t really want to spend time finding out the cause, so I reinstalled everything again today using the final installer, as well as the updated Virtualbox 4.2.6.
I wanted to try using virtuoso as the storage engine for Redland but unfortunately there is no librdf-storage-virtuoso package for Ubuntu. After getting some help from @dajobe, I attempted to build the packages myself. Although it takes quite some time to build packages, but not too difficult it seems.