Skip to content
Machine learning autonomous scripts, oriented towards ALOJA project
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
azure-ml
datasets
README.md
aloja_cli.r
deprecated.r
functions.r
models.r
precision.r
relations.r
searchtrees.r

README.md

Aloja-ML

This repo contains the scripts and tests done for the Machine Learning part of ALOJA.

Files

  • aloja_cli.r The wrapper/interface to call functions and run experiments from command line.
  • functions.r The library with all the ALOJA functions, like learning methods and data-set treatments, called from the diverse scripts and interfaces.
  • models.r Some implemented models used by functions.r.
  • precision.r Functions for computing precision of executions, also compare providers by similar executions, used by functions.r.
  • relations.r Functions for computing simple relations between features of executions, used by functions.r.
  • searchtrees.r Functions for representing execution datasets as a tree, used by functions.r.
  • deprecated.r Functions in disuse or deprecated from functions.r, kept for legacy.

Directories

  • azure-ml Version of ALOJA-ML for Microsoft AZURE-ML
  • datasets ALOJA Data-sets

Requirements

  • R and required packages: base, arules, e1071, kknn, nnet, optparse, RSNNS, rms, rpart, session, snowfall, stringr ...and dependencies.
  • build-essential

R Installation

  • Example for Ubuntu 14.04:

Shell code

wget http://security.ubuntu.com/ubuntu/pool/main/t/tiff/libtiff4\_3.9.5-2ubuntu1.6\_amd64.deb; dpkg -i ./libtiff4_3.9.5-2ubuntu1.6_amd64.deb

apt-get install "r-base" "r-base-core" "r-base-dev" "r-base-html" "r-cran-bitops" "r-cran-boot" "r-cran-class" "r-cran-cluster" "r-doc-html" "r-cran-codetools" "r-cran-foreign" "r-cran-kernsmooth" "r-cran-lattice" "r-cran-mass" "r-cran-matrix" "r-cran-mgcv" "r-cran-nlme" "r-cran-nnet" "r-cran-rpart" "r-cran-spatial" "r-cran-survival" "r-recommended" "r-cran-colorspace" "r-cran-dichromat" "r-cran-digest" "r-cran-foreach" "r-cran-gtable" "r-cran-ggplot2" "r-cran-iterators" "r-cran-labeling" "r-cran-munsell" "r-cran-plyr" "r-cran-rcolorbrewer" "r-cran-rcpp" "r-cran-reshape" "r-cran-scales" "r-cran-stringr" "gsettings-desktop-schemas" -y --force-yes

R code

install.packages(c("arules","e1071","kknn","optparse","RSNNS","rms","session","snowfall"), repos="http://cran.es.r-project.org", dependencies=TRUE,quiet=TRUE);

CLI Functionalities

Basic Syntax:

./aloja_cli.r -m method [-d dataset] [-l learned model] [-p param1=aaaa:param2=bbbb:param3=cccc:...] [-a] [-n dims] [-v]

./aloja_cli.r --method method [--dataset dataset] [--learned learned model] [--params param1=aaaa:param2=bbbb:param3=cccc:...] [--allvars] [--numvars dims] [--verbose]

Examples of Training and Prediction:

./aloja_cli.r -m aloja_regtree -d aloja-dataset.csv -p saveall=m5p1

./aloja_cli.r -m aloja_regtree -d aloja-dataset.csv -p saveall=m5p1:vin="Benchmark,Net,Disk,Maps,IO.SFac,Rep,IO.FBuf,Comp,Blk.size":vout="Exe.Time"

./aloja_cli.r -m aloja_predict_dataset -l m5p1 -d m5p1-tt.csv -v

./aloja_cli.r -m aloja_predict_instance -l m5p1 -p inst_predict="sort,ETH,RR3,8,10,1,65536,None,32,Azure L" -v

./aloja_cli.r -m aloja_predict_instance -l m5p1 -p inst_predict="sort,ETH,RR3,8|10,10,1,65536,*,32,Azure L":sorted=asc -v

./aloja_cli.r -m aloja_predict_instance -l m5p1 -p inst_predict="sort,ETH,RR3,8|10,10,1,65536,*,32,Azure L":vin="Benchmark,Net,Disk,Maps,IO.SFac,Rep,IO.FBuf,Comp, \ Blk.size,Cluster":sorted=asc -v

Examples of Detecting Outliers in the Dataset:

./aloja_cli.r -m aloja_outlier_dataset -d m5p1-tt.csv -l m5p1 -p sigma=3:hdistance=3:saveall=m5p1test

./aloja_cli.r -m aloja_outlier_instance -l m5p1 -p instance="sort,ETH,RR3,8,10,1,65536,None,32,Azure L":observed=100000:display=1 -v

Examples of Minimal Instances defining the Dataset:

./aloja_cli.r -m aloja_minimal_instances -l m5p1 -p saveall=m5p1mi

./aloja_cli.r -m aloja_minimal_instances -l m5p1 -p kmax=200:step=10:saveall=m5p1mi

Examples of JSON Tree defining the Dataset:

./aloja_cli.r -m aloja_representative_tree -p method=ordered:pred_file=instances.csv:output=string -v

Deprecated Functionalities

Examples of Dimensionality Reduction:

./aloja_cli.r -m aloja_pca -d dataset.csv -p saveall=pca1

./aloja_cli.r -m aloja_regtree -d pca1-transformed.csv -p prange=1e-4,1e+4:saveall=m5p-simple-redim -n 20

./aloja_cli.r -m aloja_predict_instance -l m5p-simple-redim -p inst_predict="1922.904354752,70.1570440421649,2.9694955079494,-3.64259027685954, \ -0.748746678239734,0.161321484374316,0.617610510007444,-0.459044093400257,0.251211132013151,0.251937462205716,-0.142007748147355,-0.0324862729758309, \ 0.406308900544488,0.13593705166432,0.397452596451088,-0.731635384355167,-0.318297127484775,-0.0876192175148721,-0.0504762335523307,-0.0146283091875174" -v

./aloja_cli.r -m aloja_predict_dataset -l m5p-simple-redim -d m5p-simple-redim-tt.csv -v

./aloja_cli.r -m aloja_transform_data -d newdataset.csv -p pca_name=pca1:saveall=newdataset

./aloja_cli.r -m aloja_transform_instance -p pca_name=pca1:inst_transform="sort,ETH,RR3,8,10,1,65536,None,32,Azure L" -v

Examples of Dataset Collapse (+Complete with prediction):

./aloja_cli.r -m aloja_dataset_collapse -d dataset.csv -p dimension1="Benchmark":dimension2="Net,Disk,Maps,IO.SFac,Rep,IO.FBuf,Comp,Blk.size, \ Cluster":dimname1="Benchmark":dimname2="Configuration":saveall=dsc1

./aloja_cli.r -m aloja_dataset_collapse -d dataset.csv -p dimension1="Benchmark":dimension2="Net,Disk,Maps,IO.SFac,Rep,IO.FBuf,Comp,Blk.size, \ Cluster":dimname1="Benchmark":dimname2="Configuration":saveall=dsc1:model_name=m5p1

./aloja_cli.r -m aloja_dataset_collapse_expand -d aloja-dataset.csv -p dimension1="Benchmark":dimension2="Net,Disk,Maps,IO.SFac,Rep,IO.FBuf, \ Comp,Blk.size,Cluster":dimname1="Benchmark":dimname2="Configuration":saveall=dsc1:model_name=m5p1:inst_general="sort,ETH,RR3,8|10,10,1,65536,*,32,Azure L"

./aloja_cli.r -m aloja_best_configurations -p bvec_name=dsc1 -v

You can’t perform that action at this time.