s434742/ium_434742
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

cutoff

Browse Source
This commit is contained in:
patrycjalazna 2021-03-28 17:21:00 +02:00
parent 6b49b51e7b
commit 6317ac1d8b
1 changed files with 7 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
avocado-preprocessing.sh
View File

@ -15,15 +15,15 @@ head -n 1 avocado-updated-2020.csv > header.csv
head -n -1 avocado-updated-2020.csv | shuf > avocado-updated-2020-shuf.csv head -n -1 avocado-updated-2020.csv | shuf > avocado-updated-2020-shuf.csv
head -n "$1" avocado-updated-2020-shuf.csv > avocado-2020.csv head -n "$1" avocado-updated-2020-shuf.csv > avocado-2020.csv
TRUNCATED_SIZE=$(wc -l avocado-2020.csv | awk '{print $1}') declare -i truncated_size=$(wc -l avocado-2020.csv | awk '{print $1}')
TEST_SIZE= $((0,2*$TRUNCATED_SIZE)) let test_size= $((0,2*$truncated_size))
DEV_SIZE=$((2*0,2*$TRUNCATED_SIZE)) let dev_size=$((2*0,2*$truncated_size))
TRAIN_SIZE=$((2*0,2*$TRUNCATED_SIZE+1)) let train_size=$((2*0,2*$truncated_size+1))
# podzial na train/dev/test 6/2/2 # podzial na train/dev/test 6/2/2
head -n $TEST_SIZE avocado-2020.csv > avocado-updated-2020-test.csv head -n $test_size avocado-2020.csv > avocado-updated-2020-test.csv
head -n $DEV_SIZE avocado-2020.csv | tail -n $TEST_SIZE > avocado-updated-2020-dev.csv head -n $dev_size avocado-2020.csv | tail -n $test_size > avocado-updated-2020-dev.csv
tail -n +$TRAIN_SIZE avocado-2020.csv > avocado-updated-2020-train.csv tail -n +$train_size avocado-2020.csv > avocado-updated-2020-train.csv
wc -l avocado-updated-2020-*.csv wc -l avocado-updated-2020-*.csv