View Code? Open in Web Editor
NEW
Stock market prediction using historical data, empirically used stock signals and / or machine learning
License: BSD 3-Clause "New" or "Revised" License
Python 1.47%
Jupyter Notebook 98.53%
stock_analyzer's Introduction
Summary of the plan:
![The plan](https://camo.githubusercontent.com/2d62c5490142425c29dd5a643405ec2b6bbca8bd0ad660c35818f32a8e5db63b/68747470733a2f2f75706c6f61642e77696b696d656469612e6f72672f77696b6970656469612f656e2f642f64642f476e6f6d65735f706c616e2e706e67)
stock_analyzer's People
Contributors
Stargazers
Watchers
stock_analyzer's Issues
- Remove old code (files / classes / functions)
- Harmonize coding style (PEP8)
-- Underscore nomenclature
-- Docstring according to google
i.e. Uwe Lang-like (Bull market, Bear market)
A starting point could be also a simple indicator, such as index majority signal.
Test of an indicator I(t) vs a baseline B(t).
Baseline could be:
- polynomial fit and extrapolation
- Fourier series
- facebook prophet extrapolation
Currently stock prices have no currency (EUR, USD, GBP, ...).
It should be obtained somehow and added to the Quotation object (or the Security object).
function add_quotation (util.py) must verify if for that day a quotation for the given symbol exists and only write if not (or overwrite).
Estimate standard deviation from past (if distribution is Gaussian...) and use this as confidence intervals around indicators.