koba-jon / normal_distribution_dataset

1-dimensional shape dataset generated by random numbers of normal distribution

Normal Distribution Dataset

1-dimensional shape dataset generated by random numbers of normal distribution

Usage

Original Dataset

Get

$ git clone https://github.com/koba-jon/normal_distribution_dataset.git
$ cd normal_distribution_dataset/NormalDistribution
$ ls -l

Hierarchy

• train : training data (100,000 pieces) of 300 dimensions

train
|--00000.dat
|--00001.dat
|  ...
|--99999.dat

• test : test data (1,000 pieces) of 300 dimensions

test
|--000.dat
|--001.dat
|  ...
|--999.dat

Create Dataset

$ sudo apt install python3 python3-pip
$ pip3 install numpy

$ vi scripts/create.sh

--dir : output directory
--num : total number of data
--dim : dimensions on one data
--list : list of mean and standard deviation of normal distribution
--seed : seed of random number

#!/bin/bash

MODE='create'

SCRIPT_DIR=$(cd $(dirname $0); pwd)

python3 ${SCRIPT_DIR}/create.py \
    --dir "./NormalDistribution/${MODE}" \
    --num 100 \
    --dim 300 \
    --list "./list/params.txt" \
    --seed 0

$ sh scripts/create.sh