Implementing k-fold crossvalidation instead of train/test split

model-py/.ipynb_checkpoints/Untitled-checkpoint.ipynb

+{
+ "cells": [],
+ "metadata": {},
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

model-py/create_cv_jobs.m

+eta = [0.01,0.05,0.1,0.2];
+max_depth = 3:10;
+nboost = 1000:1000:5000;
+
+A = combvec(eta,max_depth,nboost)';
+jobdir = '/data/user/mdefende/Projects/ctmodel-ml/model-py/cv-jobs';
+for ii = 1:length(A)
+    jobname = ['eta' num2str(A(ii,1)) '_depth' num2str(A(ii,2)) '_nboost' num2str(A(ii,3)) '.sh'];
+    
+    fid = fopen(fullfile(jobdir,jobname),'w');
+    fprintf(fid,'#!/bin/bash\n');
+    fprintf(fid,'#SBATCH --partition=pascalnodes\n');
+    fprintf(fid,'#SBATCH --time=6:00:00\n');
+    fprintf(fid,'#SBATCH --ntasks=4\n');
+    fprintf(fid,'#SBATCH --mem-per-cpu=25\n');
+    fprintf(fid,['#SBATCH --output=' jobname '.txt\n\n']);
+
+    fprintf(fid,'module load Anaconda3/5.3.0\n');
+    fprintf(fid,'source activate ctmodel-ml\n\n');
+
+    fprintf(fid,'cd /data/user/mdefende/Projects/ctmodel-ml/model-py\n\n');
+
+    fprintf(fid,['python model-indiv-cv.py ' num2str(A(ii,1)) ' ' num2str(A(ii,2)) ' ' num2str(A(ii,3)) '\n']);
+    fclose(fid);
+    
+    system(['sbatch ' fullfile(jobdir,jobname)]);
+end
+

model-py/cv-jobs/eta0.01_depth10_nboost1000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth10_nboost1000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 10 1000

model-py/cv-jobs/eta0.01_depth10_nboost2000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth10_nboost2000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 10 2000

model-py/cv-jobs/eta0.01_depth10_nboost3000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth10_nboost3000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 10 3000

model-py/cv-jobs/eta0.01_depth10_nboost4000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth10_nboost4000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 10 4000

model-py/cv-jobs/eta0.01_depth10_nboost5000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth10_nboost5000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 10 5000

model-py/cv-jobs/eta0.01_depth3_nboost1000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth3_nboost1000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 3 1000

model-py/cv-jobs/eta0.01_depth3_nboost2000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth3_nboost2000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 3 2000

model-py/cv-jobs/eta0.01_depth3_nboost3000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth3_nboost3000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 3 3000

model-py/cv-jobs/eta0.01_depth3_nboost4000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth3_nboost4000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 3 4000

model-py/cv-jobs/eta0.01_depth3_nboost5000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth3_nboost5000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 3 5000

model-py/cv-jobs/eta0.01_depth4_nboost1000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth4_nboost1000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 4 1000

model-py/cv-jobs/eta0.01_depth4_nboost2000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth4_nboost2000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 4 2000

model-py/cv-jobs/eta0.01_depth4_nboost3000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth4_nboost3000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 4 3000

model-py/cv-jobs/eta0.01_depth4_nboost4000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth4_nboost4000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 4 4000

model-py/cv-jobs/eta0.01_depth4_nboost5000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth4_nboost5000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 4 5000

model-py/cv-jobs/eta0.01_depth5_nboost1000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth5_nboost1000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 5 1000

model-py/cv-jobs/eta0.01_depth5_nboost2000.sh

+#!/bin/bash
+#SBATCH --partition=pascalnodes
+#SBATCH --time=6:00:00
+#SBATCH --ntasks=4
+#SBATCH --mem-per-cpu=25
+#SBATCH --output=eta0.01_depth5_nboost2000.sh.txt
+
+module load Anaconda3/5.3.0
+source activate ctmodel-ml
+
+cd /data/user/mdefende/Projects/ctmodel-ml/model-py
+
+python model-indiv-cv.py 0.01 5 2000