{"id":1155,"date":"2016-10-24T08:10:01","date_gmt":"2016-10-24T08:10:01","guid":{"rendered":"http:\/\/kourentzes.com\/forecasting\/?p=1155"},"modified":"2016-10-24T08:10:01","modified_gmt":"2016-10-24T08:10:01","slug":"mapa-package-for-r-on-github","status":"publish","type":"post","link":"https:\/\/kourentzes.com\/forecasting\/2016\/10\/24\/mapa-package-for-r-on-github\/","title":{"rendered":"MAPA package for R on GitHub"},"content":{"rendered":"

Here is the link: https:\/\/github.com\/trnnick\/mapa<\/a><\/p>\n

It has been a long time I wanted to rework the MAPA package for R, but I could not find the time. Finally I got around starting it. There are three objectives in this:<\/p>\n

    \n
  1. Clean up code and introduce S3methods. MAPA was the first package that I wrote!<\/li>\n
  2. Incorporate the es<\/code> exponential smoothing implementation from the smooth<\/code> package. This is an alternative to ets<\/code> that will remain as the default. It allows longer seasonal cycles and some additional customisation in the specification of the exponential smoothing models.<\/li>\n
  3. Implement MAPAx<\/a> that permits introducting exogenous variables to MAPA.<\/li>\n<\/ol>\n

    Tidying up the code is an ever-ongoing process, so (1) will never finish! I have already sorted (2) out and the next task is to work on (3). You can grab the development version on GitHub<\/a>. At the time of writing you will need the development version of smooth<\/code><\/a> for this to work, on which Ivan<\/a> has been ironing out a few peculiar bugs that MAPA brought to surface. You will be amazed how much temporal aggregation can mess up otherwise stable code!<\/p>\n

    Here is a sneak peak of things you could not do with MAPA before:<\/p>\n

    mapafit <- mapaest(taylor,ppy=336,type=\"es\")\r\nplot(mapafit)\r\n<\/pre>\n
    \"mgit_fig1\"<\/a>
    Fig. 1: Look at all these aggregation levels!<\/p><\/div>\n
    The new argument is type=\"es\"<\/code> that lets MAPA know what exponential smoothing implementation to use. Once the estimation is done we can use this to forecast as usual.<\/p>\n
    mapacalc(taylor,mapafit,comb=\"wght\",outplot=2)<\/pre>\n
    \"mgit_fig2\"<\/a>
    Fig. 2: MAPA for high frequency data!<\/p><\/div>\n
    So finally we can now easily use MAPA for series with seasonality more than 24 periods! You may have observed that I am using a different combination of the temporal aggregation states: wght<\/code>. This is still experimental, so use it at your own risk! I would still recommend using either median<\/code> or mean<\/code>.<\/p>\n
    Another nice feature that the es<\/code> implementation offers is more flexibility in specifying models. Before we could use the model=\"ZZZ\"<\/code> to let ets<\/code> select the best model for each aggregation level. Now we can use the letters X<\/code> and Y<\/code> to restrict the search to only additive or\u00a0 multiplicative options (including N<\/code>) respectively. So for example we could set model=\"ZXA\"<\/code> to have MAPA look for exponential smoothing models with any type of error, trend that can be none, additive or damped additive and additive seasonality.<\/p>\n
    Which of ets<\/code> or es<\/code> is most accurate to use with MAPA? This is a difficult question to answer. In different examples I tried, I got different results, but there are a few things I can point out. ets<\/code> from the forecast<\/code> package by default does not allow multiplicative trends. This was done as using only additive trends seems to work better for ets<\/code>. This is not the case for es<\/code> that considers all possible trends. In the current MAPA implementation you can use the argument allow.multiplicative.trend=TRUE<\/code> to make ets<\/code> consider all trends. Similarly, with es<\/code> you can use model=\"ZXZ\"<\/code> to restrict it to additive trends. On the M3 dataset, once the specification is identical (only additive trends) es<\/code> based MAPA performs about 5% worse than ets<\/code> based MAPA, which is fairly small difference, with no evidence of statistically significant differences. So my recommendation is as follows: MAPA with ets<\/code> works fine and is the default core, but if the additional flexibility of es<\/code> is needed, feel confident in using it!<\/p>\n
    Related Posts<\/H3>