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).
After the last post, I found that it may be fun to write a wrapper for YUI in order to make it behave like jQuery. Therefore, the code below is clearly mainly for self-amusement and is not intended to be used in production projects. However, through coding this, I found that although the difference in design, but YUI is obviously capable to do what jQuery offers (if not more). I will not continue working on this so whoever interested may just copy and paste the code to further developing it.
It is very much expected that there will be endless stream of new (and often times better) tools introduced to solve the same set of problems. While I am slowly resuming my programming work, and in the process of reviving my very much dead postgrad project, I found some alternative to the tools I had used in the past. I suppose I shall just jot them down here so that there’s a reference for later use.
After comparing my own implementation of MVC with CodeIgniter’s, now I’m comparing Kohana’s and Zend’s. I have just shifted from CodeIgniter to Kohana recently in work and is currently learning on how to use Zend Framework to build my web-app. As everybody knows, Zend Framework is more like a collection of library classes than a framework a la Ruby on Rails, using MVC in Zend Framework would require one to begin from bootstrapping stage. However, in Kohana, just like other frameworks, bootstrapping is done by the framework itself so the developer will get an installation that almost just works (after a little bit of configuration).
After a miserable trip back to academic world, I finally re-gained the courage to get back to job-market. For the time spent in university, I spent quite some time reading about Semantic Web and RDF. Then I thought, I should have published more in this format in future. However, that didn’t really happen, mostly because I am too lazy.
This post continued from this post. Finally I have found some time to start developing my pet project using Zend Framework. After getting the controller to work the way I am more familiar (comparing to Kohana which I used at work) with, the next step is to get it to output some data.