Skip to content

ML, DL, Data Enginnering code for Auto TA project

Notifications You must be signed in to change notification settings

cali-in-cau/auto-ta-ml

Repository files navigation

Auto-TA-ML

Authors

TBA

Summary

2021 people space group c

Prerequisites

* python3 (3.8 >= tested on 3.8.5)
* recommend making a virtual environemnt

Module Install

Actual Install

if this do not work, look at requirements.txt

pip install -r requirements.txt
# for TA-Lib
pip install TA-Lib
# https://mrjbq7.github.io/ta-lib/install.html 
# if not work, read that link

Execute

  • download_ohlc_csv.py
# excute on terminal or shell
python download_ohlc_csv.py

This will download OHLC data from yahoo finance apis, There are Ticker csv, so you can use it.

(2021-01-25) in this version you have to make daily folder to download csv data from tickers.

  • ohlc_to_image.py
# excute on terminal or shell
python ohlc_to_image.py

This will change from OHLC data(you have to download or get data first. look at download_ohlc_csv.py) to candle stick/line graph and save into png file. read the function parameter descriptions below before you excute.

  • preprocessing.py
# excute on terminal or shell
python preprocessing.py

With TAlib[https://mrjbq7.github.io/ta-lib/] we tag chart patterns automatically. and we make images with the data of OHLC, chart pattern(label). The image will contain the data with date when the chart pattern detected, including the date before and after the pattern appears.(you can change the period(function parameter))

(2021-01-25) you have to make "tmp" folder to execute. read the preprocessing.py

  • random_select_item.py
# excute on terminal or shell
python random_select_item.py [pattern_folder_name ex)2016-5years-5days-nasdaqtop300]

This file should be in the same level with the pattern_folder. Because of difference in numbers of images in each categories, it will choose same numbers of picture in each categories randomly. You can change the number of pictures.

License

TBA

TODO

  • 병렬처리 속도 향상
  • imagepid값 다시 부여
  • 폴더 재구성
  • 이미지 dropout
  • 불필요한 정보 빼기

추가해주세요

About

ML, DL, Data Enginnering code for Auto TA project

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published