/OrangeWidgets

Widgets for the Orange datamining platform

Primary LanguagePython

Jack Rogers, Tanner Robart
April 2012
New College of Florida

Goals:
	To aid scientists and academics in using orange in the study of biology
	To possibly meet satisfactory criteria for our bioinformatic machine learning
	 class without risk of failure wasting anything but time
	Learn about Orange widget creation, file parsing, metadata associations, and
	 logistic regression algorithms 

Orange Widgets:
	Bioinformatic Data Converter
		Orange is very limited by the data types it can access
		The default file widget can read data from simple 
		 tab-delimited or comma-separated files, as well as
		 files in C4.5 format and Weka’s .arrf files.		
		Bioinformatic data sets come in a variety of very unique formats that
		 store metadata in very differnt ways
		Be able to load files to data tables or convert to tab delimited
		 format
		Support for .cel, .fasta, .chp, etc... whatever orange doesnt
		 support and is convenient to implement
		
	
	Penalized Logistic Regression
		Very popular Logistic regression method over the last 5ish years in
		 regards to bioinformatic machine learning journal articles
		not present in current widget
		implemented in many other machine learning libraries
			easy to reengineer
			possibly cut and paste if credit can be given without trouble