ICOOmen_ML_Model

IcoOmen, a machine learning model which predicts the value of an ICO token after 6 months. This uses historic data that has been aggregated from various public websites and APIs, as well as data that has been manually collected and calculated.

Article can be found here.

Getting Started:

Google Colab

1. Click on the link on the top of ICOData.ipynb or click here.

Docker

1. Install Docker and Docker-Compose https://docs.docker.com/compose/install/#install-compose

2. Build and Run Docker-compose

docker-compose up

3. Head to url specified from terminal.

Jupyter-Notebooks

1. Run locally via Jupyter Notebooks.

Examples

Predict the Price of an ICO - NO TRAINING REQUIRED, USING PRE-TRAINED MODEL.

Run the following sections of Code:

• Library Imports and creating useful functions.
• Create Folders if Necessary and download dateset.
• Loading ICO dataset into variables.
• Encoding and Splitting of Data.
• Linear Regression/Neural Network.
  • Load Saved Linear Regression Models and Print out performance.
  • Use Model to make prediction - Value of ICO after 6 months.

  # Load model with best rMse and make prediction
  fileName = "results/" + "bestRegressionModel_" + str(LineaReggressionMetrics.ROOT_MEAN_SQUARED_ERROR.name) + ".sav"
  bestRegression = joblib.load(fileName)
  
  #Example ICO
  #price_usd,price_btc,total_supply,market_cap_usd,available_supply,usd_raised,eth_price_launch,btc_price_launch,ico_duration,month,day,country
  example_x = np.array([1.71456,0.00019931,1000000000,905793616,528295082,24000000,297.63,3420.4,7,8,9,182])
  
  y_pred = makePrediction(bestRegression,example_x)
  
  print("Predicted value of example ICO after 6 months: ",y_pred )

  Replace Example ICO with your ICO to predict. Use dataset/Country_Number_Mapping - Sheet1.csv to map a country to a number.