/Stock-Prediction-V3

Data Science Project

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

Earnings Momentum Strategy V3

(project in-progress)

Expanded model from https://github.com/Vxtr10/ML_stock

LSE Data Science Society Project

LSE Business Investment Group Capital - Quant Team

Week 1-2:

Statistcal Analysis on which industry earnings call has most or least effects on.

Checks mean and standard deviation of stock returns after earnings call release (day 0 being the open/close price after earnings call, whichever is first):

  • +1 day
  • +2 days
  • +7 days
  • +28 days

Week 3:

Feature Engineering and testing for correlations between different features, then selecting the best features

X_Data

(To be decided) Extracted from 3 main sources:

  1. Earnings Calls (sentiment, similarity, complexity etc.)
  2. Quarterly Reports (key word filtering etc.)
  3. Technical Analysis (MACD, Boillinger Bands etc.)

Y_Data

Stock Return between Day 0 (earnings call release date) and Day X:

  • +5 days
  • +1 week
  • +2 weeks
  • +1 month
  • +2 months