https://github.com/kenatsf/asset-fall-prediction

Model prediction
https://github.com/kenatsf/asset-fall-prediction

binomial-model bitcoin google-trends marchinelearning

Last synced: about 2 months ago
JSON representation

Model prediction

• Host: GitHub
• URL: https://github.com/kenatsf/asset-fall-prediction
• Owner: KenatSF
• Created: 2022-02-08T04:43:15.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-03-10T23:04:06.000Z (almost 3 years ago)
• Last Synced: 2024-11-10T01:14:22.786Z (about 2 months ago)
• Topics: binomial-model, bitcoin, google-trends, marchinelearning
• Language: Jupyter Notebook
• Homepage:
• Size: 5.99 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Bitcoin prediction

## Dependencies

* R version 3.6.2

* Python version 3.8.2

## Prediction

The correct prediction of any asset is almost impossible, however, one must try it.

* Making use of google trends data in words such as [Bitcoin, Blockchain, debt, dollar, ...].

* Getting Bitcoin data such as open_time, close_price.

* Getting the returns from close_price.

* Merging both data bases to built the training and testing bases.

* Comparing multiples models to get the best option.

Best model with minimum number of variables.

```

     glm(y ~ banking.system + bce + bill.gates + bitcoin.alert + 

                    bitcoin.bubble + bitcoin.dump + bitcoin.hack + bitcoin.whales + 

                    bitcoin + blockchain + coingecko + cryptonews + debt.jubilee + 

                    financial.independence + housing.market + jerome.powell + 

                    microsoft + ray.dalio + real.economy + real.estate + satoshi.nakamoto + 

                    vitalik.buterin + volatility + amazon + apple + bitcoin.twitter + 

                    ethereum + smart.contracts + sec + facebook + fed,

                data = db_training, family = "binomial")

```