topic Re: How to pass the parameter in R integrated Qlik? in Integration, Extension & APIs

How to pass the parameter in R integrated Qlik?

mohan_1105 — Mon, 03 Jul 2017 07:16:28 GMT

Hello all,

I am using the beta version of Qlik Sense. Which can be downloaded from this community https://community.qlik.com/blogs/technicalbulletin/2016/10/19/join-the-community-for-advanced-analytics-extensions

After integrating Qlik Sense with R successfully, I checked the example file TimeSeries Analysis App given by Qlik itself. In which, the example demonstrates HoltWinter forecasting passing parameter beta = FALSE and gamma = FALSE. Expression as follows,

R.ScriptEval('library(TTR);library(forecast);data.frame(forecast(HoltWinters(.arg1, beta = FALSE, gamma = FALSE), h=10))[,1]', Sum(Rain))

I tried to tweak the forecast by changing gamma = 0.1, beta = 0 and another parameter of HoltWinter function alpha = 1. The expression didn't work. Only exponential smoothing parameter(beta) works.

I am looking for the reason why it isn't working?

Re: How to pass the parameter in R integrated Qlik?

agigliotti — Mon, 03 Jul 2017 10:10:35 GMT

for what you asking for I have to look deeper inside Qlik + R integration, but now I have no time to do it.

would be great if you let us know what your app goal is and how you'll achieve it.

Re: How to pass the parameter in R integrated Qlik?

mohan_1105 — Mon, 03 Jul 2017 10:48:22 GMT

Hi Andrea,

Thank you for your time.

Consider if my data shows seasonal behavior, then I must pass Alpha(Base), beta(Smoothing) and gamma(Seasonality) parameter inside the holtwinder function for a better forecast. So I wanted to forecast this kind data with proper explanation in my app not just like that.

Any explanation on it isn't working will be much helpful.

Regards,

Mohan

Re: How to pass the parameter in R integrated Qlik?

mohan_1105 — Thu, 06 Jul 2017 09:19:09 GMT

So far, the Integration works well. There are two possible reasons for this error,

1) The forecasting data should be in the class of ts().

2) The forecasting data should have at least two years of data.

R.ScriptEval('library(TTR);library(forecast);data.frame(forecast(HoltWinters(ts(q$sumRain,start=c(1813), end=c(1912), frequency = 1), alpha = 0.598, beta=0, gamma=0.904), h=$(RainNumberOfYears)))[,1]', Sum(rain) as sumRain)