search

michael-spengler / decentralized-finance

Distributed Ledger based Banking Features: Payments, Deposits, Loans and Automated investment Patterns
MIT License
39 stars 5 forks source link

[Feature] Improve the automated investment pattern. #1

Open michael-spengler opened 3 years ago

michael-spengler commented 3 years ago

Prelude The following proposals shall be understood considering the "net cash on average" having the reinvestment rounds or redemption rounds already in the equation.

Optimize the Degree of the Leverage Effect It seems to be a good idea to automatically and slowly decrease the leverage effect along times of rising asset prices in order to protect the investment during times of falling asset prices. According to that it is probably reasonable to automatically and slowly increase the leverage effect along times of falling asset prices in order to increase the probability of generating a strong outcome.

In other words The health factor of an investment should increase during rising asset prices It is probably also reasonable that it decreases during falling asset prices

Implementation Shall be done in the following class: https://github.com/michael-spengler/decentralized-finance/blob/main/src/thug-life-investments/thug-life.service.ts

michael-spengler commented 3 years ago

Just an additional optional pattern: Combine investing with artificial neural net based trading. The neural net shall leverage the following kinds of training data:

  1. wale wallet transactions
  2. sentiment analysis
  3. classical trader stuff like deriving future price development guesses from past prices and price correlations (https://www.investopedia.com/terms/a/atr.asp...)

Explore and leverage the following technological assets:

  1. https://www.npmjs.com/package/brain.js
  2. https://www.tensorflow.org/js
  3. https://deno.land/x/neural_nets@0.0.3
  4. https://www.npmjs.com/package/node-nlp (for the sentiment analysis)