Welcome to the New Site!

Chris Woodall | 14 Aug 2013 | | news

Welcome to the new static BUILDS website powered by Bootstrap 2.3.2 and Jekyll. This website was designed by Chris Woodall and will provide you with event updates, project updates and more!

Also, if you want to you can fork this website on Github and push us updates! We will be vetting pushes and the officers will be the primary maintainers of this website.


Older · View Archive (146)

Tackle a Kaggle

Tackle a Kaggle with BUILDS! Kaggle is a machine learning competition platform

Overview

Our first competition:

   The sinking of the RMS Titanic is one of the most infamous shipwrecks in history.  On April 15, 1912, 
during her maiden voyage, the Titanic sank after colliding with an iceberg, killing 1502 out of 2224 
passengers and crew.  This sensational tragedy shocked the international community and led to better safety 
regulations for ships.
   One of the reasons that the shipwreck led to such loss of life was that there were not enough lifeboats 
for the passengers and crew.  Although there was some element of luck involved in surviving the sinking, 
some groups of people were more likely to survive than others, such as women, children, and the upper-class.
   In this contest, we ask you to complete the analysis of what sorts of people were likely to survive. In
particular, we ask you to apply the tools of machine learning to predict which passengers survived the tragedy.

Previous Meetings

Meeting Times: Saturdays at 1500 EST

2013-05-11

  • Preliminary meeting
  • Gained familiarity with python

2013-05-18

  • Tested homebrew code against sample code
  • Made multiple submissions
  • Discussed possibility of different models

2013-05-25

  • Digits Recognizer competition
  • completed one-vs-rest linear SVM, which placed 1121/1176 (83% accuracy)
  • benchmark of KNN placed 350/1176 (96% accuracy)

2013-06-01

  • implemented SVM with RBF kernel and tested (worse than benchmark)

2013-06-08

  • had lots of fun

2013-06-15

  • Started Linear Models with Amazon competition~

2013-06-22

  • prepared for HAMR 2013 (logistics, etc)

2013-06-29

  • HAMR 2013

2013-07-06

  • no meeting

2013-07-13

  • no meeting

2013-07-20

2013-07-29

  • Convolutional Neural Networks tutorial (takes a REALLY long time to run)
  • asked permission for csa servers to install necessary libraries because ^

Leaderboard Results

Titanic: 733/2734

Digits Recognizer: 350/1176

Newer

Meeting Minutes for 2013-09-04

Notes recorded from BUILDS meeting Recorded 9/04/13