Crypto-Data-Labeller

Uses triple barrier method to label data. Contains a checker to inform the profitability of a trading bot.

What is triple barrier method labelling?

The idea behind the triple-barrier method is that we have three barriers: an upper barrier, a lower barrier, and a vertical barrier. The upper barrier represents the threshold an observation’s return needs to reach in order to be considered a buying opportunity (a label of 1), the lower barrier represents the threshold an observation’s return needs to reach in order to be considered a selling opportunity (a label of -1), and the vertical barrier represents the amount of time an observation has to reach its given return in either direction before it is given a label of 0.

How to use the labeller script above?

1. Install all the required packages and import them in python
2. Convert your own dataset to parquet format (Saves time and is more space-efficient)
3. Change the path in the labeller script to your own dataset
4. Run labeller-slow through backtesting.py (This will give more insight but takes a long time)
5. Run labeller-fast for fast coversion of dataset into labels

How to use the checker script ?

1. Create a "product.csv" with ohlcv and labels(assuming the labels are from your own trading bot)

Open,High,Low,Close,Volume,labels
46368.73,46391.49,46314.26,46331.08,30.45894,0
46331.07,46336.1,46300.0,46321.34,20.96029,0
46321.34,46443.56,46280.0,46436.03,35.86682,0
46436.03,46518.32,46432.5,46518.31,29.31849,0
46518.31,46527.26,46427.06,46427.95,27.81847,0

2. You can use the y.csv file to check the labels distribution

labels.to_csv('data/y.csv')
print(labels.value_counts())
labels.value_counts().plot.bar()

3. Run the backtest function to get the metrics spreadsheet and you can easily evaluate your trading bot's performance

Dependencies

• pandas
• backtesting.py
• numpy