fHMM/README.Rmd at master · timoadam/fHMM · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
---
output: github_document
---

<!-- README.md is generated from README.Rmd. Please edit that file -->

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "75%",
  fig.align = "center"
)
set.seed(1)
library(fHMM)
```

# {fHMM}: Fitting HMMs to financial time series <img src="man/figures/logo.svg" align="right" alt="" width="120" />

<!-- badges: start -->
[![R-CMD-check](https://github.com/loelschlaeger/fHMM/workflows/R-CMD-check/badge.svg)](https://github.com/loelschlaeger/fHMM/actions)
[![CRAN status](https://www.r-pkg.org/badges/version-last-release/fHMM)](https://www.r-pkg.org/badges/version-last-release/fHMM)
[![CRAN downloads](https://cranlogs.r-pkg.org/badges/grand-total/fHMM)](https://cranlogs.r-pkg.org/badges/grand-total/fHMM)
[![codecov](https://codecov.io/gh/loelschlaeger/fHMM/branch/master/graph/badge.svg)](https://app.codecov.io/gh/loelschlaeger/fHMM)
<!-- badges: end -->

With {fHMM} you can detect and characterize financial market regimes in financial time series by applying hidden Markov Models (HMMs). The functionality and the model [is documented in detail here](https://loelschlaeger.de/fHMM/articles/). Below, you can find a first application to the German stock index [DAX](https://en.wikipedia.org/wiki/DAX).

## Installation

You can install the released version of {fHMM} from [CRAN](https://CRAN.R-project.org) with:

```{r, eval = FALSE}
install.packages("fHMM")
```

And the development version from [GitHub](https://github.com/) with:

```{r, eval = FALSE}
# install.packages("devtools")
devtools::install_github("loelschlaeger/fHMM")
```

## Contributing

We welcome contributions! Please submit [bug reports](https://github.com/loelschlaeger/fHMM/issues/new?assignees=&labels=bug&template=bug.md) and [feature requests](https://github.com/loelschlaeger/fHMM/issues/new?assignees=&labels=future&template=suggestion.md) as issues and extensions as pull request from a branch forked from main.

## Example: Fitting an HMM to the DAX

We fit a 2-state HMM with state-dependent t-distributions to the DAX log-returns from 2000 to 2020. The states can be interpreted as proxies for bearish and bullish markets.

The package has a build-in function to download the data from [Yahoo Finance](https://finance.yahoo.com/):

```{r data download}
path <- paste0(tempdir(),"/dax.csv")
download_data(symbol = "^GDAXI", file = path, verbose = FALSE)
```

We first need to define the model by setting some `controls`:

```{r controls}
controls <- list(
  states = 2,
  sdds   = "t",
  data   = list(file        = path,
                date_column = "Date",
                data_column = "Close",
                logreturns  = TRUE,
                from        = "2000-01-01",
                to          = "2020-12-31")
)
controls <- set_controls(controls)
```

The function `prepare_data()` prepares the data for estimation:

```{r data}
data <- prepare_data(controls)
summary(data)
```

We fit the model and subsequently decode the hidden states:

```{r fit}
model <- fit_model(data, ncluster = 7)
model <- decode_states(model)
summary(model)
```

Having estimated the model, we can visualize the state-dependent distributions and the decoded time series:

```{r plots, fig.dim = c(10,6), fig.align='center'}
events <- fHMM_events(
  list(dates = c("2001-09-11", "2008-09-15", "2020-01-27"),
       labels = c("9/11 terrorist attack", "Bankruptcy Lehman Brothers", "First COVID-19 case Germany"))
)
plot(model, plot_type = c("sdds","ts"), events = events)
```