eagle-kaist 0.0.1

Stock Prediction Pipeline

Step 1 - Build Stock Database

Step 2 - Normalize Stock Data

Step 3 - Build Features and Visualization

Step 3 - Build Prediction Models

Metrics

The metrics of prediction models should be based on profits and loss they give.

1. Profit vs. loss

2. Backtest

Stock data

Collected stock data include following values:

• Close
• Open
• High
• Low
• Volumn
• Change
• Percentage Change

Derivative values:

• Return
• Alpha
• Volatility
• Alpha*
• Sharpe

Business Taxonomy

There is a large number of companies in stock market. It is crucial to take a look over their financial reports and market indices to have a correct view of the companies' business status.

In this section, we want to build several main groups of companies. Each group represents the area of business that the company is taking part in. In addition, we target to give a common credit score for each group, and also the rank of company in that group.

• Input:

  • Stock indices
  • Companies financial reports

• Output:

  • We want to have a comparison between the areas and give the advice on which business we should invest in.
  • We want to have a table that ranks companies and shows their potential in market.

Installation:

conda create --name eagle python=3.7
conda activate eagle
pip install -r requirements.txt