Financial machine learning

The goal of the fml package is to learning about the emerging field of financial machine learning and its applications in computer age econometrics. This package contains templates for reports, and functions and workshops using in Algorithmic trading and investment) taught by Barry Quinn in Queen’s Management School.

The package os inspired by the fantastic work of Marcos López de Prado, and his books entitled Advances in Financial Machine Learning and Machine learning for Asset Managers.

Installation

Install/or reinstall the package from GitHub using the following.

remove.packages('fml')
.rs.restartR()
remotes::install_github("quinfer/fml")

Example

This is a basic example which shows you how to solve a common problem:

library(fml)
library(tidyverse)
#> ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
#> ✓ ggplot2 3.3.5     ✓ purrr   0.3.4
#> ✓ tibble  3.1.6     ✓ dplyr   1.0.7
#> ✓ tidyr   1.1.4     ✓ stringr 1.4.0
#> ✓ readr   2.1.1     ✓ forcats 0.5.1
#> ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
#> x dplyr::filter() masks stats::filter()
#> x dplyr::lag()    masks stats::lag()
## basic example code
fml::daily_factors %>% summary()
#>       date                  rm                   rf           
#>  Min.   :1988-10-03   Min.   :-0.0834130   Min.   :0.0000000  
#>  1st Qu.:1996-01-23   1st Qu.:-0.0046057   1st Qu.:0.0000210  
#>  Median :2003-05-19   Median : 0.0006790   Median :0.0001825  
#>  Mean   :2003-05-15   Mean   : 0.0003942   Mean   :0.0001765  
#>  3rd Qu.:2010-09-06   3rd Qu.: 0.0055839   3rd Qu.:0.0002231  
#>  Max.   :2017-12-29   Max.   : 0.0921039   Max.   :0.0005430  
#>       rmrf                 smb                  hml            
#>  Min.   :-0.0835837   Min.   :-6.301e-02   Min.   :-4.187e-02  
#>  1st Qu.:-0.0047718   1st Qu.:-3.755e-03   1st Qu.:-2.963e-03  
#>  Median : 0.0005000   Median : 1.062e-04   Median :-7.734e-05  
#>  Mean   : 0.0002178   Mean   :-2.793e-05   Mean   : 6.069e-05  
#>  3rd Qu.: 0.0053843   3rd Qu.: 3.883e-03   3rd Qu.: 2.888e-03  
#>  Max.   : 0.0920217   Max.   : 3.561e-02   Max.   : 5.784e-02  
#>       umd            
#>  Min.   :-0.0813362  
#>  1st Qu.:-0.0034154  
#>  Median :-0.0001573  
#>  Mean   :-0.0003336  
#>  3rd Qu.: 0.0030267  
#>  Max.   : 0.0599399

Function test

?fml::estRMT()

Tutorials

The tutorials can be run on a local machine only. You can start the tutorials in one of two ways. First, in RStudio 1.3 or later, you will find the ATI tutorials listed in the “Tutorial” tab in the top-right pane (by default). Find a tutorial and click “Run Tutorial” to get started. Second, you can run any tutorial from the R console by typing the following line:

learnr::run_tutorial("Workshop2","fml")

This should bring up a tutorial in your default web browser. You can see the full list of tutorials by running:

learnr::run_tutorial(package = "fml")

Critical Essay

This package also includes a RMarkdown template for use in the critical essay assessment. Go to File>New>R Markdown… and choose from From Template then Report.

Datasets

FTSE 350 data

The package includes point in time FTSE350 data from 2016-2020, downloaded from Refinitiv Datastream for teaching purposes only. The data has been used to create two return series 1. A point in time Top 25 by average market value returns series 2. A current Top 30 by market capitalisation returns series

fml::ftse350
fml::ftse30_returns_mthly
fml::ftse25_rtns_mthly

Daily uk asset pricing risk factors

These are created by Essex university business school and downloaded from UK Data Service API.

fml::daily_factors"