https://github.com/krzjoa/aeon

Time Series with keras
https://github.com/krzjoa/aeon

Last synced: 5 days ago
JSON representation

Time Series with keras

• Host: GitHub
• URL: https://github.com/krzjoa/aeon
• Owner: krzjoa
• License: other
• Created: 2022-05-30T18:54:52.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-06-26T07:05:48.000Z (over 1 year ago)
• Last Synced: 2024-11-07T15:07:40.895Z (about 2 months ago)
• Language: R
• Homepage: https://krzjoa.github.io/aeon
• Size: 9.68 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.Rmd
  • Funding: .github/FUNDING.yml
  • License: LICENSE

Awesome Lists containing this project

README

        
---

output: github_document

---

```{r, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>",

  fig.path = "man/figures/README-",

  out.width = "100%"

)

```

# aion



[![CRAN status](https://www.r-pkg.org/badges/version/aion)](https://CRAN.R-project.org/package=aion)

[![Buy hex

stciker](https://img.shields.io/badge/buy%20hex-aion-green?style=flat&logo=redbubble)](https://www.redbubble.com/i/sticker/aion-R-package-hex-by-krzjoa/126321293.EJUG5)

> Time Series models for keras in R 

## Installation

You can install the development version of aion from [GitHub](https://github.com/) with:

``` r

# install.packages("devtools")

devtools::install_github("krzjoa/aion")

```

## Key features

* Temporal Fusion Transformer model

* additional layers: Temporal Convolutional Network block & Legendre Memory Unit

* `make_array` and `ts_generator` functions to quickly prepare input/output for `{keras}`

time series models

* new loss functions: `loss_quantile`, `loss_tweedie` and `loss_negative_log_likelihood`

## Usage

```{r}

# Dataset

library(m5)

# Neural Networks

library(aion)

library(keras)

# Data wrangling

library(dplyr, warn.conflicts=FALSE)

library(data.table, warn.conflicts=FALSE)

library(recipes, warn.conflicts=FALSE)

# ==========================================================================

#                          PREPARING THE DATA

# ==========================================================================

train <- tiny_m5[date < '2016-01-01']

test  <- tiny_m5[date >= '2016-01-01']

m5_recipe <-

  recipe(value ~ ., data=train) %>%

  step_mutate(item_id_idx=item_id, store_id_idx=store_id) %>%

  step_integer(item_id_idx, store_id_idx,

               wday, month,

               event_name_1, event_type_1,

               event_name_2, event_type_2,

               zero_based=TRUE) %>%

  step_naomit(all_predictors()) %>%

  prep()

train <- bake(m5_recipe, train)

test  <- bake(m5_recipe, test)

TARGET      <- 'value'

STATIC      <- c('item_id_idx', 'store_id_idx')

CATEGORICAL <- c('event_name_1', 'event_type_1', STATIC)

NUMERIC     <- c('sell_price', 'sell_price')

KEY         <- c('item_id', 'store_id')

INDEX       <- 'date'

LOOKBACK    <- 28

HORIZON     <- 14

STRIDE      <- LOOKBACK

BATCH_SIZE  <- 32

# ==========================================================================

#                          CREATING GENERATORS

# ==========================================================================

c(train_generator, train_steps) %<-%

    ts_generator(

        data        = train,

        key         = KEY,

        index       = INDEX,

        lookback    = LOOKBACK,

        horizon     = HORIZON,

        stride      = STRIDE,

        target      = TARGET,

        static      = STATIC,

        categorical = CATEGORICAL,

        numeric     = NUMERIC,

        batch_size  = BATCH_SIZE    

  )

c(test_generator, test_steps)  %<-%

    ts_generator(

        data = test,

        key = KEY,

        index = INDEX,

        lookback = LOOKBACK,

        horizon = HORIZON,

        stride = STRIDE,

        target=TARGET,

        static=STATIC,

        categorical=CATEGORICAL,

        numeric=NUMERIC

    )

```