dmc - DATA-MINING-CUP-2012 project

Age	Commit message (Collapse)	Author	Files	Lines
2012-05-14	doc: WIP presentation	Jan Klemkow	1	-7/+19

2012-05-14	Add distances to frequency removal regression slide	Benjamin Franzke	1	-0/+1

2012-05-14	regress: Add an ugly workaround hack for old octave versions	Benjamin Franzke	1	-2/+11

2012-05-14	Dont return negative quantities in regression	Benjamin Franzke	2	-0/+2

2012-05-13	doc: Describing some methods	Jan Klemkow	3	-18/+72

2012-05-12	Add a interval based regression	Benjamin Franzke	2	-0/+27
	But its actually a very ugly prediction.
2012-05-12	Explicitly state that price indexes have to correlate with quantities	Benjamin Franzke	1	-3/+3

2012-05-12	regress_pred: Add degree param, to fold in quad regression	Benjamin Franzke	5	-29/+10

2012-05-12	Fold log_regress_pred into regress_pred	Benjamin Franzke	4	-21/+6
	By adding a pre-filter function pointer.
2012-05-12	Use polyval where appropriate	Benjamin Franzke	4	-20/+8
	..instead of doing the polynom multiplication by hand.
2012-05-12	Add classification result to .gitignore	Benjamin Franzke	1	-0/+1

2012-05-12	Implement calc_error and total_error as anonymous functions	Benjamin Franzke	3	-10/+3

2012-05-12	get_products: Use better variable names	Benjamin Franzke	1	-3/+3

2012-05-12	get_products: Use reshape to generate the day by product matrix	Benjamin Franzke	1	-14/+2

2012-05-12	classify: Print classified table to file	Benjamin Franzke	1	-1/+13

2012-05-12	Add classification script	Benjamin Franzke	1	-0/+12

2012-05-12	run_tests: Drop random prediction method	Jan Klemkow	1	-1/+1
	Verified by: rnd_prod = (method_idx == 6); run_tests; rnd_prod = (rnd_prod & (method_idx == 6)); sum(rnd_prod)
2012-05-12	get_products: Return the real price classification set as well	Benjamin Franzke	2	-2/+10

2012-05-12	Add classification data in octave readable format	Benjamin Franzke	1	-0/+7980

2012-05-12	Add sed script that converts data into octave readable csv files	Benjamin Franzke	1	-0/+4
	Usage: ./gen_octave_input.sed dmc2012_task/train.txt > train_noheader.csv
2012-05-12	opt_pred: Split opt_pred in opt_idx and opt_pred.	Jan Klemkow	3	-10/+14
	So you can use the idx of opt_idx for other predictions.
2012-05-12	Add dmc 2012 official task files	Benjamin Franzke	5	-0/+32030

2012-05-12	Add .gitignore	Benjamin Franzke	1	-0/+6

2012-05-12	doc: Add ploting scripts for sevenday and random.	Jan Klemkow	2	-0/+32

2012-05-12	chi_pred: add an inactive chi rand. pred method.	Jan Klemkow	2	-1/+11

2012-05-12	opt_pred: preallocate matrices.	Jan Klemkow	1	-5/+4

2012-05-11	run_tests: Disable all plots	Benjamin Franzke	2	-8/+8

2012-05-11	rand_pred: Pred needs to have 570 columns, not 1	Benjamin Franzke	1	-2/+2

2012-05-11	praesentation: Fix my email	Benjamin Franzke	1	-1/+1

2012-05-11	doc: wip	Jan Klemkow	1	-0/+2

2012-05-11	doc: add presentation	Jan Klemkow	2	-54/+39

2012-05-11	graph: change style and code	Jan Klemkow	1	-1/+1

2012-05-11	rand_pred: Code clean up.	Jan Klemkow	1	-3/+3

2012-05-11	run_tests: Pack struct definition into a anonymous function	Benjamin Franzke	1	-7/+8

2012-05-11	run_tests: Use end-keyword for appending to a vector	Benjamin Franzke	1	-1/+1

2012-05-11	get_products: Force the use of ascii	Benjamin Franzke	1	-1/+1
	Just to be sure.
2012-05-11	graph: Add tango colors map and script for chats.	Jan Klemkow	2	-0/+56

2012-05-11	run_tests: Use an vector of structs for the prediction methods	Benjamin Franzke	1	-11/+11
	Allows use to use e.g. pred_methods.name.
2012-05-11	run_tests: Optimize local_min_count storage	Benjamin Franzke	1	-2/+2

2012-05-11	run_tests: Add optimze errors into qerr/terr directly	Benjamin Franzke	1	-3/+1

2012-05-11	Put local error calculation into calc_error	Benjamin Franzke	2	-7/+5

2012-05-11	Store all non-emta prediction methods in a cell-array	Benjamin Franzke	1	-30/+25

2012-05-11	Silence regress_frequency_removal output	Benjamin Franzke	1	-2/+2

2012-05-11	regress_pred: No need to embed regress	Benjamin Franzke	1	-6/+0

2012-05-11	Move frequency removal regression into its own function	Benjamin Franzke	2	-7/+16

2012-05-11	Remove log+quad algorithm	Benjamin Franzke	1	-4/+1
	Its really not needed since wehave optimize
2012-05-11	sevenday_pred: Use same prototype as for other tests	Benjamin Franzke	2	-6/+13

2012-05-11	mean/rand_pred: Use same function prototype as regress_pred	Benjamin Franzke	3	-4/+6
	The number of days is implicit in the size of the given price and quantity matrices.
2012-05-11	normalize: Normalize max=min to 1	Benjamin Franzke	1	-2/+4

2012-05-11	run_tests: Store local_min_count in a variable as well	Benjamin Franzke	1	-2/+4