mewo2 / medley Goto Github PK
View Code? Open in Web Editor NEWA system for blending regression models in R
License: BSD 2-Clause "Simplified" License
A system for blending regression models in R
License: BSD 2-Clause "Simplified" License
To reproduce the problem please run this script:
require(medley);
require(e1071);
data(swiss)
x <- swiss[,1:5];
y <- swiss[,6];
set.seed(1)
train <- sample(nrow(swiss), 30);
m <- create.medley(x[train,], y[train]);
for (gamma in c(1e-3, 2e-3, 5e-3, 1e-2, 2e-2, 5e-2, 1e-1)) {
m <- add.medley(m, svm, list(gamma=gamma));
}
# use only the best 80% of individual models
m <- prune.medley(m, 0.8)
p <- predict(m, x[-train,]);
rmse(p, y[-train]);
I have the next output:
CV model 1 svm (gamma = 0.001) time: 0.03 error: 2.649394
CV model 2 svm (gamma = 0.002) time: 0.02 error: 2.645305
CV model 3 svm (gamma = 0.005) time: 0.03 error: 2.615573
CV model 4 svm (gamma = 0.01) time: 0.04 error: 2.607492
CV model 5 svm (gamma = 0.02) time: 0.03 error: 2.599948
CV model 6 svm (gamma = 0.05) time: 0.03 error: 2.646888
CV model 7 svm (gamma = 0.1) time: 0.04 error: 2.604546
Sampled...
Error in apply(mixpred, 1, mixer) : dim(X) must have a positive length
Error in mean((true - pred)^2) : object 'p' not found
At the moment, the ensemble is calculated at prediction time. This is obviously inefficient if more than one predict() call is made. It also means that it's impossible to interrogate the ensemble object to determine which models were used, etc.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.