Skip to content

Instantly share code, notes, and snippets.

Avatar
🚲

Cyril Auburtin caub

🚲
View GitHub Profile
@caub
caub / KLRS.md
Last active Aug 29, 2015
Kernel Recursive Least Square using JSAT https://code.google.com/p/java-statistical-analysis-tool/, and test on Santa Fe laser data
View KLRS.md

result

NMSE = 0.039
@caub
caub / NNet.md
Last active Aug 29, 2015
Neural Net test on Santa Fe laser data with code.google.com/p/java-statistical-analysis-tool and sourceforge.net/projects/jarbm
View NNet.md

RBMNet

nmse = 0.095
@caub
caub / Http-test.mq4
Last active Aug 29, 2015
Small MT4 (http://docs.mql4.com/) script for orders copying between 2 accounts
View Http-test.mq4
#include <http51.mqh>
extern string url = "http://localhost:8080/orders";
//this scripts sends an http post containing orders info each tick to this url
int start() {
string params [0,2];
int status[1]; // HTTP Status code
int total=OrdersTotal();
View santa-fe.r
require(nnet)
require(caret)
y = read.csv('http://www-psych.stanford.edu/~andreas/Time-Series/SantaFe/A.dat', header=F)
y2 = read.csv('http://www-psych.stanford.edu/~andreas/Time-Series/SantaFe/A.cont', header=F)
k = 40
n=100
y = y$V1/256
y2 = y2$V1/256
dat = sapply(1:k, function(a) c(rep(NA,a),y[1:(length(y)-a)]) )
@caub
caub / KRLS plot.md
Last active Aug 29, 2015
Santa-Fe regression with matlab
View KRLS plot.md

krls

nmse = 0.042

@caub
caub / knapsack01.m
Last active Aug 29, 2015
an interview question
View knapsack01.m
% problem: we have a result, and many elements, find all combinations that sum up to the result
function sol = knapsack01(maxCapacity, items)
% knapsack problem with variables in {0,1}
% Naive solution is O(n!), knapsack implementation is O(n*m) where n is
% items length and m is weights length
@caub
caub / Zigzag.md
Last active Aug 29, 2015
zigzag chart indicator, a piecewise linear curve fit with alternate slopes (up, down, up, down...), used here to detect double-top and double-bottom patterns http://en.wikipedia.org/wiki/Chart_pattern
View rsi-bt.py
from scipy import *
M = 22 #ma for rsi
N = 14 #rsi loopback
thresh = [20,80] #rsi thresholds
cost = 0.0001 # cost per trade (spread)
price = 1.3 + 0.1*randn(100) + sin(linspace(0,10,100))
ma = ema(price, M)
ri = rsindex(price-ma, N)
@caub
caub / svm.py
Last active Aug 29, 2015
Simplified SMO from coursera's ml-class converted from matlab
View svm.py
from scipy import *
#from scipy.linalg import *
from pylab import *
class SVM:
def train(self, X, Y, kernel, C, tol = 1e-3, max_passes = 5):
m = size(X, 0)
n = size(X, 1)
@caub
caub / polyMultiFeatures.m
Created Apr 28, 2014
Combinatorics for generating polynomial terms of degree <=k
View polyMultiFeatures.m
function [ as_ ] = polyMultiFeatures( items, k )
as = [];
function recurse(a, i)
% we should optimize and early stop a with length>k
if i>size(items,2)
if size(a,2)<=k
as{end+1} = a;
end
return;
end